Giovanni Sileno, profile picture
Uploaded byGiovanni Sileno
79 views

Code-driven Law NO, Normware SI!

The document discusses the distinctions between various legal frameworks, specifically code-driven, data-driven, and text-driven laws, highlighting their implications for automated decision-making in the legal field. It emphasizes the challenges of ensuring legal certainty and justice when laws are articulated in computer code and the necessity for feedback mechanisms to address potential failures. The text suggests that while computational legalism and justice ideals present opportunities, they also risk oversimplification of complex legal norms and concepts.

Embed presentation

Download to read offline
Code-driven Law NO, Normware SI! 3 November 2022, CRCL conference “Computational Law on Edge”, Brussels Giovanni Sileno, g.sileno@uva.nl Socially Intelligent Artificial Systems (SIAS), Informatics Institute, University of Amsterdam
Part I: types of “law”, and a more nuanced distinction
text-driven law Hildebrandt, M., Code Driven Law. Scaling the Past and Freezing the Future, In: Is Law Computable? Critical Perspectives on Law and Artificial Intelligence, eds. Markou, Deakin, Hart Publishers (2020) code-driven law data-driven law
text-driven law legal activity performed by humans by means of sources of norms such as statute and case law Hildebrandt, M., Code Driven Law. Scaling the Past and Freezing the Future, In: Is Law Computable? Critical Perspectives on Law and Artificial Intelligence, eds. Markou, Deakin, Hart Publishers (2020) code-driven law data-driven law
data-driven law text-driven law legal activity performed by humans by means of sources of norms such as statute and case law automatic decision-making or predictions used in support derived from statistical/inductive methods Hildebrandt, M., Code Driven Law. Scaling the Past and Freezing the Future, In: Is Law Computable? Critical Perspectives on Law and Artificial Intelligence, eds. Markou, Deakin, Hart Publishers (2020) code-driven law
code-driven law data-driven law text-driven law legal norms or policies that have been articulated in computer code legal activity performed by humans by means of sources of norms such as statute and case law automatic decision-making or predictions used in support derived from statistical/inductive methods Hildebrandt, M., Code Driven Law. Scaling the Past and Freezing the Future, In: Is Law Computable? Critical Perspectives on Law and Artificial Intelligence, eds. Markou, Deakin, Hart Publishers (2020)
code-driven law data-driven law text-driven law Hildebrandt, M., Code Driven Law. Scaling the Past and Freezing the Future, In: Is Law Computable? Critical Perspectives on Law and Artificial Intelligence, eds. Markou, Deakin, Hart Publishers (2020) legal activity performed by humans by means of sources of norms such as statute and case law open-textured concepts multi-interpretability → it can be contested legal norms or policies that have been articulated in computer code automatic decision-making or predictions used in support derived from statistical/inductive methods
code-driven law data-driven law text-driven law Hildebrandt, M., Code Driven Law. Scaling the Past and Freezing the Future, In: Is Law Computable? Critical Perspectives on Law and Artificial Intelligence, eds. Markou, Deakin, Hart Publishers (2020) legal activity performed by humans by means of sources of norms such as statute and case law open-textured concepts multi-interpretability → it can be contested statistical closure logical closure legal norms or policies that have been articulated in computer code automatic decision-making or predictions used in support derived from statistical/inductive methods
code-driven law data-driven law text-driven law Hildebrandt, M., Code Driven Law. Scaling the Past and Freezing the Future, In: Is Law Computable? Critical Perspectives on Law and Artificial Intelligence, eds. Markou, Deakin, Hart Publishers (2020) legal activity performed by humans by means of sources of norms such as statute and case law open-textured concepts multi-interpretability → it can be contested statistical closure logical closure legal certainty + legal norms or policies that have been articulated in computer code automatic decision-making or predictions used in support derived from statistical/inductive methods
code-driven law data-driven law text-driven law Hildebrandt, M., Code Driven Law. Scaling the Past and Freezing the Future, In: Is Law Computable? Critical Perspectives on Law and Artificial Intelligence, eds. Markou, Deakin, Hart Publishers (2020) legal activity performed by humans by means of sources of norms such as statute and case law open-textured concepts multi-interpretability → it can be contested statistical closure logical closure legal certainty + justice (proportionality, distribution, …) ? instrumentality (contextual policy goals) ? legal norms or policies that have been articulated in computer code automatic decision-making or predictions used in support derived from statistical/inductive methods
code-driven law data-driven law text-driven law Hildebrandt, M., Code Driven Law. Scaling the Past and Freezing the Future, In: Is Law Computable? Critical Perspectives on Law and Artificial Intelligence, eds. Markou, Deakin, Hart Publishers (2020) legal activity performed by humans by means of sources of norms such as statute and case law open-textured concepts multi-interpretability → it can be contested statistical closure logical closure COMPUTATIONAL LEGALISM legal certainty + justice (proportionality, distribution, …) ? instrumentality (contextual policy goals) ? legal norms or policies that have been articulated in computer code automatic decision-making or predictions used in support derived from statistical/inductive methods
code-driven law data-driven law text-driven law legal activity performed by humans by means of sources of norms such as statute and case law legal norms or policies that have been articulated in computer code automatic decision-making or predictions used in support derived from statistical/inductive methods statistical closure logical closure legal certainty + justice (proportionality, distribution, …) ? instrumentality (contextual policy goals) ?
code-driven law data-driven law text-guided law legal activity performed by humans by means of sources of norms such as statute and case law decision support systems statistical closure logical closure legal certainty + justice (proportionality, distribution, …) ? instrumentality (contextual policy goals) ? in so far humans are involved data-guided law automatic decision-making derived from statistical/inductive methods legal norms or policies that have been articulated in computer code
code-driven law data-driven law text-guided law legal activity performed by humans by means of sources of norms such as statute and case law decision support systems statistical closure logical closure legal certainty + justice (proportionality, distribution, …) ? instrumentality (contextual policy goals) ? data-guided law automatic decision-making derived from statistical/inductive methods instance + label instance + label instance + label instance + label ….. TRAINING DATASET legal norms or policies that have been articulated in computer code
code-driven law data-driven law text-guided law legal activity performed by humans by means of sources of norms such as statute and case law decision support systems statistical closure logical closure legal certainty + justice (proportionality, distribution, …) ? instrumentality (contextual policy goals) ? data-guided law automatic decision-making derived from statistical/inductive methods instance + label instance + label instance + label instance + label ….. is the training data set representative of the domain in which it will be applied? is the rationality behind labeling still the same? TRAINING DATASET legal norms or policies that have been articulated in computer code
code-driven law data-driven law text-guided law legal activity performed by humans by means of sources of norms such as statute and case law decision support systems statistical closure logical closure legal certainty + justice (proportionality, distribution, …) ? instrumentality (contextual policy goals) ? data-guided law automatic decision-making derived from statistical/inductive methods instance + label instance + label instance + label instance + label ….. is the training data set representative of the domain in which it will be applied? is the rationality behind labeling still the same? TRAINING DATASET is the training method producing adequate, reproducible, robust results? TRAINING METHOD legal norms or policies that have been articulated in computer code
code-driven law data-driven law text-guided law legal activity performed by humans by means of sources of norms such as statute and case law decision support systems statistical closure logical closure legal certainty + justice (proportionality, distribution, …) ? instrumentality (contextual policy goals) ? data-guided law automatic decision-making derived from statistical/inductive methods instance + label instance + label instance + label instance + label ….. is the training data set representative of the domain in which it will be applied? is the rationality behind labeling still the same? TRAINING DATASET is the training method producing adequate, reproducible, robust results? TRAINING METHOD → it can be contested legal norms or policies that have been articulated in computer code
code-driven law data-driven law text-guided law legal activity performed by humans by means of sources of norms such as statute and case law decision support systems statistical closure logical closure legal certainty + justice (proportionality, distribution, …) ? instrumentality (contextual policy goals) ? data-guided law automatic decision-making derived from statistical/inductive methods fact fact …. STRUCTURAL rule rule ….. CONTINGENT legal norms or policies that have been articulated in computer code
code-driven law data-driven law text-guided law legal activity performed by humans by means of sources of norms such as statute and case law decision support systems statistical closure logical closure legal certainty + justice (proportionality, distribution, …) ? instrumentality (contextual policy goals) ? data-guided law automatic decision-making derived from statistical/inductive methods fact fact …. STRUCTURAL rule rule ….. CONTINGENT were all the relevant rules considered? were all the relevant facts given? is the formalization correct? is the reasoning tool robust? legal norms or policies that have been articulated in computer code
code-driven law data-driven law text-guided law legal activity performed by humans by means of sources of norms such as statute and case law decision support systems statistical closure logical closure legal certainty + justice (proportionality, distribution, …) ? instrumentality (contextual policy goals) ? data-guided law automatic decision-making derived from statistical/inductive methods fact fact …. STRUCTURAL rule rule ….. CONTINGENT → it can be contested were all the relevant rules considered? were all the relevant facts given? is the formalization correct? is the reasoning tool robust? legal norms or policies that have been articulated in computer code
code-driven law data-driven law text-guided law legal activity performed by humans by means of sources of norms such as statute and case law decision support systems statistical closure logical closure legal certainty ? justice (proportionality, distribution, …) ? instrumentality (contextual policy goals) ? data-guided law automatic decision-making derived from statistical/inductive methods the COMPUTATIONAL LEGALISM “ideal” works only in so far we forget all what may go wrong legal norms or policies that have been articulated in computer code
code-driven law data-driven law text-guided law legal activity performed by humans by means of sources of norms such as statute and case law decision support systems statistical closure logical closure legal certainty ? justice (proportionality, distribution, …) ? instrumentality (contextual policy goals) ? data-guided law automatic decision-making derived from statistical/inductive methods is this all? legal norms or policies that have been articulated in computer code
code-driven law data-driven law text-guided law legal activity performed by humans by means of sources of norms such as statute and case law decision support systems statistical closure logical closure legal certainty ? justice (proportionality, distribution, …) + instrumentality (contextual policy goals) ? data-guided law automatic decision-making derived from statistical/inductive methods these principles could be to some extent formalized as meta-rules legal norms or policies that have been articulated in computer code
code-driven law data-driven law text-guided law legal activity performed by humans by means of sources of norms such as statute and case law decision support systems statistical closure logical closure legal certainty ? justice (proportionality, distribution, …) + instrumentality (contextual policy goals) + data-guided law automatic decision-making derived from statistical/inductive methods these principles could be to some extent formalized as meta-rules these goals could be made explicit legal norms or policies that have been articulated in computer code
code-driven law data-driven law text-guided law legal activity performed by humans by means of sources of norms such as statute and case law decision support systems statistical closure logical closure legal certainty ? justice (proportionality, distribution, …) + instrumentality (contextual policy goals) + data-guided law automatic decision-making derived from statistical/inductive methods these principles could be to some extent formalized as meta-rules these goals could be made explicit can we be sure about a certain formalization? legal norms or policies that have been articulated in computer code
code-driven law data-driven law text-guided law legal activity performed by humans by means of sources of norms such as statute and case law decision support systems statistical closure logical closure legal certainty ? justice (proportionality, distribution, …) ? instrumentality (contextual policy goals) + data-guided law automatic decision-making derived from statistical/inductive methods these principles could be to some extent formalized as meta-rules these goals could be made explicit never. can we be sure about a certain formalization? (cf. discussions on universal human rights, ethical frameworks) legal norms or policies that have been articulated in computer code
code-driven law data-driven law text-guided law legal activity performed by humans by means of sources of norms such as statute and case law decision support systems statistical closure logical closure legal certainty ? justice (proportionality, distribution, …) ? instrumentality (contextual policy goals) + data-guided law automatic decision-making derived from statistical/inductive methods these principles could be to some extent formalized as meta-rules these goals could be made explicit may we be ever ready for that? (explicit articulation of policy-maker purposes) never. can we be sure about a certain formalization? (cf. discussions on universal human rights, ethical frameworks) legal norms or policies that have been articulated in computer code
code-driven law data-driven law text-guided law legal activity performed by humans by means of sources of norms such as statute and case law decision support systems statistical closure logical closure legal certainty ? justice (proportionality, distribution, …) ? instrumentality (contextual policy goals) ? data-guided law automatic decision-making derived from statistical/inductive methods these principles could be to some extent formalized as meta-rules these goals could be made explicit may we be ever ready for that? (explicit articulation of policy-maker purposes) not to a full extent. never. can we be sure about a certain formalization? (cf. discussions on universal human rights, ethical frameworks) (it may require to be transparent on highly contestable issues) legal norms or policies that have been articulated in computer code
code-driven law data-driven law text-guided law legal activity performed by humans by means of sources of norms such as statute and case law decision support systems statistical closure logical closure legal certainty ? justice (proportionality, distribution, …) ? instrumentality (contextual policy goals) ? data-guided law automatic decision-making derived from statistical/inductive methods these principles could be to some extent formalized as meta-rules these goals could be made explicit may we be ever ready for that? (explicit articulation of policy-maker purposes) not to a full extent. never. can we be sure about a certain formalization? (cf. discussions on universal human rights, ethical frameworks) (it may require to be transparent on highly contestable issues) the COMPUTATIONAL JUSTICE “ideal” assumes perfect “conceptualization” legal norms or policies that have been articulated in computer code
code-driven law data-driven law text-guided law legal activity performed by humans by means of sources of norms such as statute and case law decision support systems legal certainty ? justice (proportionality, distribution) ? instrumentality (contextual policy goals) ? data-guided law automatic decision-making derived from statistical/inductive methods the COMPUTATIONAL JUSTICE “ideal” assumes perfect “conceptualization” the COMPUTATIONAL LEGALISM “ideal” works only in so far we forget all what may go wrong legal norms or policies that have been articulated in computer code
code-driven law data-driven law text-guided law legal activity performed by humans by means of sources of norms such as statute and case law decision support systems legal certainty ? justice (proportionality, distribution) ? instrumentality (contextual policy goals) ? data-guided law automatic decision-making derived from statistical/inductive methods the COMPUTATIONAL JUSTICE “ideal” assumes perfect “conceptualization” the COMPUTATIONAL LEGALISM “ideal” works only in so far we forget all what may go wrong let’s accept things may go wrong at all level legal norms or policies that have been articulated in computer code
code-driven law data-driven law text-guided law legal activity performed by humans by means of sources of norms such as statute and case law decision support systems legal certainty ? justice (proportionality, distribution) ? instrumentality (contextual policy goals) ? data-guided law automatic decision-making derived from statistical/inductive methods the COMPUTATIONAL JUSTICE “ideal” assumes perfect “conceptualization” the COMPUTATIONAL LEGALISM “ideal” works only in so far we forget all what may go wrong let’s accept things may go wrong at all level people or other actors should be able to appeal legal norms or policies that have been articulated in computer code
code-driven law data-driven law text-guided law legal activity performed by humans by means of sources of norms such as statute and case law decision support systems legal certainty ? justice (proportionality, distribution) ? instrumentality (contextual policy goals) ? data-guided law automatic decision-making derived from statistical/inductive methods the COMPUTATIONAL JUSTICE “ideal” assumes perfect “conceptualization” the COMPUTATIONAL LEGALISM “ideal” works only in so far we forget all what may go wrong let’s accept things may go wrong at all level let’s accept we need to integrate systematically feedbacks from higher-order courts, jurisprudence or other normative sources people or other actors should be able to appeal legal norms or policies that have been articulated in computer code
code-driven law data-driven law text-guided law legal activity performed by humans by means of sources of norms such as statute and case law decision support systems legal certainty ? justice (proportionality, distribution) ? instrumentality (contextual policy goals) ? data-guided law automatic decision-making derived from statistical/inductive methods let’s accept things may go wrong at all level let’s accept we need to integrate systematically feedbacks from higher-order courts, jurisprudence or other normative sources people or other actors should be able to appeal courts or other actors should be able to quash legal norms or policies that have been articulated in computer code the COMPUTATIONAL JUSTICE “ideal” assumes perfect “conceptualization” the COMPUTATIONAL LEGALISM “ideal” works only in so far we forget all what may go wrong
including computational actors possibility of continuous, automated testing/verification code-driven law data-driven law text-guided law legal activity performed by humans by means of sources of norms such as statute and case law decision support systems legal certainty ? justice (proportionality, distribution) ? instrumentality (contextual policy goals) ? data-guided law automatic decision-making derived from statistical/inductive methods let’s accept things may go wrong at all level let’s accept we need to integrate systematically feedbacks from higher-order courts, jurisprudence or other normative sources people or other actors should be able to appeal courts or other actors should be able to quash legal norms or policies that have been articulated in computer code the COMPUTATIONAL JUSTICE “ideal” assumes perfect “conceptualization” the COMPUTATIONAL LEGALISM “ideal” works only in so far we forget all what may go wrong
including computational actors possibility of continuous, automated testing/verification code-driven law data-driven law text-guided law derived from legal norms or policies that have been articulated in computer code (or from other methods) lower authority w.r.t. human authorities!!! legal activity performed by humans by means of sources of norms such as statute and case law decision support systems legal certainty ? justice (proportionality, distribution) ? instrumentality (contextual policy goals) ? data-guided law automatic decision-making derived from statistical/inductive methods the COMPUTATIONAL JUSTICE “ideal” assumes perfect “conceptualization” the COMPUTATIONAL LEGALISM “ideal” works only in so far we forget all what may go wrong let’s accept things may go wrong at all level let’s accept we need to integrate systematically feedbacks from higher-order courts, jurisprudence or other normative sources people or other actors should be able to appeal courts or other actors should be able to quash code-guided law monolithical systems ecological system including interfaces with humans
including computational actors possibility of continuous, automated testing/verification code-driven law data-driven law text-guided law derived from legal norms or policies that have been articulated in computer code (or from other methods) lower authority w.r.t. human authorities!!! legal activity performed by humans by means of sources of norms such as statute and case law decision support systems legal certainty ? justice (proportionality, distribution) ? instrumentality (contextual policy goals) ? data-guided law automatic decision-making derived from statistical/inductive methods the COMPUTATIONAL JUSTICE “ideal” assumes perfect “conceptualization” the COMPUTATIONAL LEGALISM “ideal” works only in so far we forget all what may go wrong let’s accept things may go wrong at all level let’s accept we need to integrate systematically feedbacks from higher-order courts, jurisprudence or other normative sources people or other actors should be able to appeal courts or other actors should be able to quash code-guided law monolithical systems ecological system including interfaces with humans a serious technological gap exists today...
Part II: what is normware?
A tentative ontology SOFTWARE HARDWARE ● physical device ● when running ⇒ physical process ● situated in a physical environment ● symbolic device ● when running ⇒ symbolic process ● relies on physical processes
? A tentative ontology SOFTWARE HARDWARE NORMWARE ● physical device ● when running ⇒ physical process ● situated in a physical environment ● symbolic device ● when running ⇒ symbolic process ● relies on physical processes ● ? ● when running ⇒ ? ● relies on ?
? A tentative ontology SOFTWARE HARDWARE NORMWARE ● physical device ● when running ⇒ physical process ● situated in a physical environment ● symbolic device ● when running ⇒ symbolic process ● relies on physical processes ● ? ● when running ⇒ ? ● relies on ? ARTIFACT dimension PROCESS dimension
Normware as artifacts: directives concerning regulation the cookie jar must be full you are prohibited to eat cookies you can not eat cookies
Normware as artifacts: directives concerning regulation aiming to regulate behaviour aiming to regulate situations in the world the cookie jar must be full you are prohibited to eat cookies you can not eat cookies
we can have explicit normative specifications… Normware as artifacts: directives concerning regulation access-control policies languages based on deontic logic languages based on Hohfeld’s primitives
but programs in themselves are already mandatory in nature! a := 2 + 2 system has to perform 2 + 2… ?mother(maggie, bart) system has to prove that… animal :- dog. system has to make animal true if dog is true Normware as artifacts: directives concerning regulation
Normware as artifacts: directives concerning terminology the cookie jar must be full you are prohibited to eat cookies you can not eat cookies what is a cookie? what does it mean to be full? what is eating? who is you? what is a jar?
Normware as artifacts: directives concerning expectations the cookie jar must be full you are prohibited to eat cookies you can not eat cookies eating cookies → cookies are destroyed → the jar is not full practical normative reasoning always require some world knowledge
Normware as artifacts: devices intended to regulate doors regulate entrances semaphores regulate traffic
Normware as artifacts: devices intended to regulate doors regulate entrances semaphores regulate traffic we do not have access to the inner decision-making mechanism that brings to the current output state!
Normware as artifacts: devices intended to regulate black-boxes (eg. ML models) are also artifacts expressing some form of normativity/normality “is this a cat?” “how to (best) behave in certain conditions?”
Normware as artifacts: devices intended to regulate black-boxes (eg. ML models) are also artifacts expressing some form of normativity/normality “is this a cat?” “how to (best) behave in certain conditions?” from a functional point of view, they also count as normware!
Whether artificial or natural, designed or emergent, what counts in control is ● the existence of some reference (the target of control), ● which the entity is set to either approach or avoid (the direction of control). Normware as processes: regulation as control
Whether artificial or natural, designed or emergent, what counts in control is ● the existence of some reference (the target of control), ● which the entity is set to either approach or avoid (the direction of control). behavioural or situational space intervention current state target state Normware as processes: regulation as control
what defines the references though? target state current state direction Normware as processes: regulation as control
Whether artificial or natural, designed or emergent, what counts in control is ● the existence of some reference (the target of control), ● which the entity is set to either approach or avoid (the direction of control). by defining directives by this control signature (target, direction), any regulative mechanisms can be abstracted from its implementation. direction Normware as processes: regulation as control current state target state
indeterminacy of references Normware as processes: higher-order indetermination
indeterminacy of references indeterminacy of directives Normware as processes: higher-order indetermination
indeterminacy of references antinomies indeterminacy of directives Normware as processes: higher-order indetermination
indeterminacy of references antinomies indeterminacy of directives Normware as processes: higher-order indetermination mechanisms of conflict resolution are needed at systematic level!
Normware: first-order control GOAL ACTION
Normware: first-order control GOAL ACTION labelled dataset ML black box instruction operation data-driven code-driven
Normware: first-order control GOAL ACTION labelled dataset ML black box instruction operation data-driven code-driven pool of actions In both cases, some method has been used to select the action from a pool of actions…
GOAL ACTION But then, where the goal comes from? pool of actions Normware: first-order control labelled dataset ML black box instruction operation data-driven code-driven
TACTICAL GOAL ACTION pool of actions Normware: second-order control STRATEGIC GOAL Adding depth!
pool of goals TACTICAL GOAL ACTION pool of actions Normware: second-order control STRATEGIC GOAL Adding depth!
pool of goals TACTICAL GOAL ACTION pool of actions Normware: second-order control STRATEGIC GOAL Sileno, G., Boer, A. and van Engers, T., The Role of Normware in Trustworthy and Explainable AI, Proceedings of XAILA workshop: Explainable AI and Law, in conjunction with JURIX 2018 trustworthy AI and explainable AI issues in ML due to lack of the strategic component Adding depth!
TACTICAL GOAL ACTION Normware: second-order control STRATEGIC GOAL cybernetic view on systems: policy, intelligence, operations
TACTICAL GOAL ACTION Normware: second-order control STRATEGIC GOAL …yet it is about a single “organism”, not an “ecology” “TOTALITARIAN” architecture cybernetic view on systems: policy, intelligence, operations
Normware: plural second-order control we need to acknowledge the presence of several autonomous entities,
Normware: plural second-order control we need to acknowledge the presence of several autonomous entities,
Normware: plural second-order control we need to acknowledge the presence of several autonomous entities, and adequate conflict resolution mechanisms
Normware: plural second-order control resolution mechanisms may also result from a similar process
Normware: plural second-order control resolution mechanisms may also result from a similar process “PARTICIPATORY” architecture
Normware: plural second-order control resolution mechanisms may also result from a similar process “PARTICIPATORY” architecture Sileno, G., Grosso, P., Accounting Value Effects for Responsible Networking. Proceedings of ACM SIGCOMM 2021 Workshop on Technologies, Applications, and Uses of a Responsible Internet (TAURIN2021).
? SOFTWARE HARDWARE NORMWARE physical device when running ⇒ physical process situated in a physical environment symbolic device when running ⇒ symbolic process relies on physical processes coordination device when running ⇒ coordination process relies on symbolic (possibly hard-coded) processes A less tentative ontology
Key points ● There is continuity between institutional and computational activities. ● We don’t need institutions to become more mechanical, institutional constructs need to be brought into the computational realm. ● As we have a plurality of normative sources, we need a plurality of computational normative sources. normware
mechanizing law introducing legitimate normative processes within the computational realm SI! NO Contemporary socio-technical challenge:
Contemporary socio-technical challenge: mechanizing law introducing legitimate normative processes within the computational realm SI! code-driven law data-driven law normware law-makers, policy-makers, citizens, … NO
Code-driven Law NO, Normware SI! 3 November 2022, CRCL conference “Computational Law on Edge”, Brussels Giovanni Sileno, g.sileno@uva.nl Socially Intelligent Artificial Systems (SIAS), Informatics Institute, University of Amsterdam

More Related Content

PDF
Code Driven Law?
byGiovanni Sileno
 
PDF
Computational Legal Studies The Promise And Challenge Of Datadriven Research ...
byvukeladayski43
 
PDF
Creating Accessible Online Learning for Students with Disabilities
bykaydeverka
 
PPTX
Artificial Intelligence and Law
bySamos2019Summit
 
PDF
Nicolas Petit 27 september 18 - Hard Questions of Law and AI
byNicolas Petit
 
PDF
Access to justice through virtual doors - Daniela Piana
byOECD Governance
 
PPTX
Machine-consumable legislation - Waddington
byMatthewWaddington3
 
PDF
On Mapping Values in AI Governance
byGiovanni Sileno
 
Code Driven Law?
byGiovanni Sileno
 
Computational Legal Studies The Promise And Challenge Of Datadriven Research ...
byvukeladayski43
 
Creating Accessible Online Learning for Students with Disabilities
bykaydeverka
 
Artificial Intelligence and Law
bySamos2019Summit
 
Nicolas Petit 27 september 18 - Hard Questions of Law and AI
byNicolas Petit
 
Access to justice through virtual doors - Daniela Piana
byOECD Governance
 
Machine-consumable legislation - Waddington
byMatthewWaddington3
 
On Mapping Values in AI Governance
byGiovanni Sileno
 

Similar to Code-driven Law NO, Normware SI!

PPTX
(DAMPS 2013) E-services via the Internet and compliance with the law. File 20...
byVytautas Čyras
 
PPTX
Open Legal Data Workshop at Stanford
byHarry Surden
 
PDF
AI for Legal Research with applications, tools
bymahaffeycheryld
 
PDF
AI use cases in legal research - An Overview.pdf
byJamieDornan2
 
PDF
Legal Markup Generation in the Large: An Experience Report
byLionel Briand
 
PPT
jjjjjjjjjjjjjjjjjjjjjjjjjjjjjjjjjjjjjjjjjjjjjjjjjjjjjjj.ppt
byADRIKOPATRICIA
 
PDF
Addressing the challenges of harmonizing law and artificial intelligence tech...
byIAESIJAI
 
PDF
Introduction to Legal Technology, lecture 2 (2015)
byAnna Ronkainen
 
PDF
Technology, Data and Computation Session @ The World Bank - Law, Justice, and...
byDaniel Katz
 
PDF
A new era of efficiency and accuracy.pdf
byChristopherTHyatt
 
PDF
Automated Extraction of Semantic Legal Metadata Using Natural Language Proces...
byLionel Briand
 
PDF
Iv 4 Sp 12 Presentation 001
byIESS
 
PDF
What is legal technology?
byAnna Ronkainen
 
PPTX
Legal Informatics Research Today: Implications for Legal Prediction, 3D Print...
byRobert Richards
 
PDF
Ai and applications in the legal domain studium generale maastricht 20191101
byjcscholtes
 
PDF
AI FOR LEGAL RESEARCH: STREAMLINING LEGAL PRACTICES FOR THE DIGITAL AGE
byChristopherTHyatt
 
PDF
Law + Complexity & Prediction: Toward a Characterization of Legal Systems as ...
byDaniel Katz
 
PDF
Research Handbook On Law And Technology Bartosz Broek Olia Kanevskaia
byinisaniyang
 
PDF
Approaches To Legal Ontologies Theories Domains Methodologies 1st Edition Pom...
bytbwcjpcrj094
 
PDF
Machine learning and the challenges of digital transformation in the law
bySebastian Ko
 
(DAMPS 2013) E-services via the Internet and compliance with the law. File 20...
byVytautas Čyras
 
Open Legal Data Workshop at Stanford
byHarry Surden
 
AI for Legal Research with applications, tools
bymahaffeycheryld
 
AI use cases in legal research - An Overview.pdf
byJamieDornan2
 
Legal Markup Generation in the Large: An Experience Report
byLionel Briand
 
jjjjjjjjjjjjjjjjjjjjjjjjjjjjjjjjjjjjjjjjjjjjjjjjjjjjjjj.ppt
byADRIKOPATRICIA
 
Addressing the challenges of harmonizing law and artificial intelligence tech...
byIAESIJAI
 
Introduction to Legal Technology, lecture 2 (2015)
byAnna Ronkainen
 
Technology, Data and Computation Session @ The World Bank - Law, Justice, and...
byDaniel Katz
 
A new era of efficiency and accuracy.pdf
byChristopherTHyatt
 
Automated Extraction of Semantic Legal Metadata Using Natural Language Proces...
byLionel Briand
 
Iv 4 Sp 12 Presentation 001
byIESS
 
What is legal technology?
byAnna Ronkainen
 
Legal Informatics Research Today: Implications for Legal Prediction, 3D Print...
byRobert Richards
 
Ai and applications in the legal domain studium generale maastricht 20191101
byjcscholtes
 
AI FOR LEGAL RESEARCH: STREAMLINING LEGAL PRACTICES FOR THE DIGITAL AGE
byChristopherTHyatt
 
Law + Complexity & Prediction: Toward a Characterization of Legal Systems as ...
byDaniel Katz
 
Research Handbook On Law And Technology Bartosz Broek Olia Kanevskaia
byinisaniyang
 
Approaches To Legal Ontologies Theories Domains Methodologies 1st Edition Pom...
bytbwcjpcrj094
 
Machine learning and the challenges of digital transformation in the law
bySebastian Ko
 

More from Giovanni Sileno

PDF
DPCL: a Language Template for Normative Specifications
byGiovanni Sileno
 
PDF
Unexpectedness and Bayes' Rule
byGiovanni Sileno
 
PDF
Accounting Value Effects for Responsible Networking
byGiovanni Sileno
 
PDF
Operationalizing Declarative and Procedural Knowledge
byGiovanni Sileno
 
PDF
History of AI, Current Trends, Prospective Trajectories
byGiovanni Sileno
 
PDF
The Role of Normware in Trustworthy and Explainable AI
byGiovanni Sileno
 
PDF
Computing Contrast on Conceptual Spaces
byGiovanni Sileno
 
PDF
On the problems of interface: explainability, conceptual spaces, relevance
byGiovanni Sileno
 
PDF
A Computational Model of Moral and Legal Responsibility via Simplicity Theory
byGiovanni Sileno
 
PDF
Aligning Law and Action
byGiovanni Sileno
 
PDF
Similarity and Contrast on Conceptual Spaces for Pertinent Description Genera...
byGiovanni Sileno
 
PDF
A Petri net-based notation for normative modeling: evaluation on deontic para...
byGiovanni Sileno
 
PDF
Reading Agendas Between the Lines, an exercise
byGiovanni Sileno
 
PDF
Bridging Representation of Laws, of Implementations and of Behaviours
byGiovanni Sileno
 
PDF
Revisiting Constitutive Rules
byGiovanni Sileno
 
PDF
Commitments, Expectations, Affordances and Susceptibilities: Towards Position...
byGiovanni Sileno
 
PDF
A Constructivist Approach to Rule Bases
byGiovanni Sileno
 
PDF
On the Interactional Meaning of Fundamental Legal Concepts
byGiovanni Sileno
 
PPTX
Legal Knowledge Conveyed by Narratives: towards a representational model
byGiovanni Sileno
 
PDF
Implementing Explanation-Based Argumentation using Answer Set Programming
byGiovanni Sileno
 
DPCL: a Language Template for Normative Specifications
byGiovanni Sileno
 
Unexpectedness and Bayes' Rule
byGiovanni Sileno
 
Accounting Value Effects for Responsible Networking
byGiovanni Sileno
 
Operationalizing Declarative and Procedural Knowledge
byGiovanni Sileno
 
History of AI, Current Trends, Prospective Trajectories
byGiovanni Sileno
 
The Role of Normware in Trustworthy and Explainable AI
byGiovanni Sileno
 
Computing Contrast on Conceptual Spaces
byGiovanni Sileno
 
On the problems of interface: explainability, conceptual spaces, relevance
byGiovanni Sileno
 
A Computational Model of Moral and Legal Responsibility via Simplicity Theory
byGiovanni Sileno
 
Aligning Law and Action
byGiovanni Sileno
 
Similarity and Contrast on Conceptual Spaces for Pertinent Description Genera...
byGiovanni Sileno
 
A Petri net-based notation for normative modeling: evaluation on deontic para...
byGiovanni Sileno
 
Reading Agendas Between the Lines, an exercise
byGiovanni Sileno
 
Bridging Representation of Laws, of Implementations and of Behaviours
byGiovanni Sileno
 
Revisiting Constitutive Rules
byGiovanni Sileno
 
Commitments, Expectations, Affordances and Susceptibilities: Towards Position...
byGiovanni Sileno
 
A Constructivist Approach to Rule Bases
byGiovanni Sileno
 
On the Interactional Meaning of Fundamental Legal Concepts
byGiovanni Sileno
 
Legal Knowledge Conveyed by Narratives: towards a representational model
byGiovanni Sileno
 
Implementing Explanation-Based Argumentation using Answer Set Programming
byGiovanni Sileno
 

Recently uploaded

PPTX
Diffusion Models under Local Differential Privacy for Personalized Human Mobi...
bybl4150
 
PPTX
How to Select the Right CMMS Software for Your Organization — A Complete Buye...
byMaintWiz Technologies Private Limited
 
PPTX
TPM Metrics & Measurement: Drive Performance Excellence with TPM | MaintWiz
byMaintWiz Technologies Private Limited
 
PPTX
UnrealGameplayAbilitySystemPresentation.pptx
byBenHowenstein
 
PPTX
Natural Gas fundamentals and GRU for associated gas trap.pptx
byengineerhassan
 
PPTX
Importance of Instrument Calibration — Improve Accuracy, Reliability & Mainte...
byMaintWiz Technologies Private Limited
 
PPTX
Lean Maintenance: Streamlining Operations for Efficiency & Reliability
byMaintWiz Technologies Private Limited
 
PPTX
firewall Selection in production life pptx
byssuserb1479b
 
PDF
Narrows Planning Collective Transportation Capstone.pdf
bycj2392
 
PDF
FABRICATION AND CHARACTERIZATION OF CONDUCTING POLYMER COMPOSITE
byijoejnl
 
PPTX
Basic Electrical Engineering-Introduction to Light Sources
byKhuman Haobam
 
PPTX
How to Create an Effective Monthly Maintenance Plan for Reliable Plant Operat...
byMaintWiz Technologies Private Limited
 
PDF
Albert Pintoy - Specializing In Low-Latency
byAlbert Pintoy
 
PPTX
Avoiding Pitfalls: The 7 Biggest Maintenance Mistakes Plant Managers Make
byMaintWiz Technologies Private Limited
 
PDF
Lecture -06-Hybrid Policies - Chapter 7- Weeks 6-7.pdf
bysalaarmike
 
PPTX
Preventive Maintenance for Material Handling Equipment — Complete Guide
byMaintWiz Technologies Private Limited
 
PDF
ACI 318-2205_American Concrete Institute.pdf
byericaguilerasoto1
 
PPT
63490613-Boiler-Tube-Leakage-analysis-symptoms-causes.ppt
byallahdittashafi
 
PPTX
Step-by-step guide to designing standard a microbiology laboratory in pharmac...
byshawkyabdo1
 
PDF
CV_Rahul_ELV Project Engineer_8 Years Experience
byRahulSushruth
 
Diffusion Models under Local Differential Privacy for Personalized Human Mobi...
bybl4150
 
How to Select the Right CMMS Software for Your Organization — A Complete Buye...
byMaintWiz Technologies Private Limited
 
TPM Metrics & Measurement: Drive Performance Excellence with TPM | MaintWiz
byMaintWiz Technologies Private Limited
 
UnrealGameplayAbilitySystemPresentation.pptx
byBenHowenstein
 
Natural Gas fundamentals and GRU for associated gas trap.pptx
byengineerhassan
 
Importance of Instrument Calibration — Improve Accuracy, Reliability & Mainte...
byMaintWiz Technologies Private Limited
 
Lean Maintenance: Streamlining Operations for Efficiency & Reliability
byMaintWiz Technologies Private Limited
 
firewall Selection in production life pptx
byssuserb1479b
 
Narrows Planning Collective Transportation Capstone.pdf
bycj2392
 
FABRICATION AND CHARACTERIZATION OF CONDUCTING POLYMER COMPOSITE
byijoejnl
 
Basic Electrical Engineering-Introduction to Light Sources
byKhuman Haobam
 
How to Create an Effective Monthly Maintenance Plan for Reliable Plant Operat...
byMaintWiz Technologies Private Limited
 
Albert Pintoy - Specializing In Low-Latency
byAlbert Pintoy
 
Avoiding Pitfalls: The 7 Biggest Maintenance Mistakes Plant Managers Make
byMaintWiz Technologies Private Limited
 
Lecture -06-Hybrid Policies - Chapter 7- Weeks 6-7.pdf
bysalaarmike
 
Preventive Maintenance for Material Handling Equipment — Complete Guide
byMaintWiz Technologies Private Limited
 
ACI 318-2205_American Concrete Institute.pdf
byericaguilerasoto1
 
63490613-Boiler-Tube-Leakage-analysis-symptoms-causes.ppt
byallahdittashafi
 
Step-by-step guide to designing standard a microbiology laboratory in pharmac...
byshawkyabdo1
 
CV_Rahul_ELV Project Engineer_8 Years Experience
byRahulSushruth
 

Code-driven Law NO, Normware SI!

  • 1.
    Code-driven Law NO, NormwareSI! 3 November 2022, CRCL conference “Computational Law on Edge”, Brussels Giovanni Sileno, g.sileno@uva.nl Socially Intelligent Artificial Systems (SIAS), Informatics Institute, University of Amsterdam
  • 2.
    Part I: types of“law”, and a more nuanced distinction
  • 3.
    text-driven law Hildebrandt, M.,Code Driven Law. Scaling the Past and Freezing the Future, In: Is Law Computable? Critical Perspectives on Law and Artificial Intelligence, eds. Markou, Deakin, Hart Publishers (2020) code-driven law data-driven law
  • 4.
    text-driven law legal activityperformed by humans by means of sources of norms such as statute and case law Hildebrandt, M., Code Driven Law. Scaling the Past and Freezing the Future, In: Is Law Computable? Critical Perspectives on Law and Artificial Intelligence, eds. Markou, Deakin, Hart Publishers (2020) code-driven law data-driven law
  • 5.
    data-driven law text-driven law legalactivity performed by humans by means of sources of norms such as statute and case law automatic decision-making or predictions used in support derived from statistical/inductive methods Hildebrandt, M., Code Driven Law. Scaling the Past and Freezing the Future, In: Is Law Computable? Critical Perspectives on Law and Artificial Intelligence, eds. Markou, Deakin, Hart Publishers (2020) code-driven law
  • 6.
    code-driven law data-driven law text-drivenlaw legal norms or policies that have been articulated in computer code legal activity performed by humans by means of sources of norms such as statute and case law automatic decision-making or predictions used in support derived from statistical/inductive methods Hildebrandt, M., Code Driven Law. Scaling the Past and Freezing the Future, In: Is Law Computable? Critical Perspectives on Law and Artificial Intelligence, eds. Markou, Deakin, Hart Publishers (2020)
  • 7.
    code-driven law data-driven law text-drivenlaw Hildebrandt, M., Code Driven Law. Scaling the Past and Freezing the Future, In: Is Law Computable? Critical Perspectives on Law and Artificial Intelligence, eds. Markou, Deakin, Hart Publishers (2020) legal activity performed by humans by means of sources of norms such as statute and case law open-textured concepts multi-interpretability → it can be contested legal norms or policies that have been articulated in computer code automatic decision-making or predictions used in support derived from statistical/inductive methods
  • 8.
    code-driven law data-driven law text-drivenlaw Hildebrandt, M., Code Driven Law. Scaling the Past and Freezing the Future, In: Is Law Computable? Critical Perspectives on Law and Artificial Intelligence, eds. Markou, Deakin, Hart Publishers (2020) legal activity performed by humans by means of sources of norms such as statute and case law open-textured concepts multi-interpretability → it can be contested statistical closure logical closure legal norms or policies that have been articulated in computer code automatic decision-making or predictions used in support derived from statistical/inductive methods
  • 9.
    code-driven law data-driven law text-drivenlaw Hildebrandt, M., Code Driven Law. Scaling the Past and Freezing the Future, In: Is Law Computable? Critical Perspectives on Law and Artificial Intelligence, eds. Markou, Deakin, Hart Publishers (2020) legal activity performed by humans by means of sources of norms such as statute and case law open-textured concepts multi-interpretability → it can be contested statistical closure logical closure legal certainty + legal norms or policies that have been articulated in computer code automatic decision-making or predictions used in support derived from statistical/inductive methods
  • 10.
    code-driven law data-driven law text-drivenlaw Hildebrandt, M., Code Driven Law. Scaling the Past and Freezing the Future, In: Is Law Computable? Critical Perspectives on Law and Artificial Intelligence, eds. Markou, Deakin, Hart Publishers (2020) legal activity performed by humans by means of sources of norms such as statute and case law open-textured concepts multi-interpretability → it can be contested statistical closure logical closure legal certainty + justice (proportionality, distribution, …) ? instrumentality (contextual policy goals) ? legal norms or policies that have been articulated in computer code automatic decision-making or predictions used in support derived from statistical/inductive methods
  • 11.
    code-driven law data-driven law text-drivenlaw Hildebrandt, M., Code Driven Law. Scaling the Past and Freezing the Future, In: Is Law Computable? Critical Perspectives on Law and Artificial Intelligence, eds. Markou, Deakin, Hart Publishers (2020) legal activity performed by humans by means of sources of norms such as statute and case law open-textured concepts multi-interpretability → it can be contested statistical closure logical closure COMPUTATIONAL LEGALISM legal certainty + justice (proportionality, distribution, …) ? instrumentality (contextual policy goals) ? legal norms or policies that have been articulated in computer code automatic decision-making or predictions used in support derived from statistical/inductive methods
  • 12.
    code-driven law data-driven law text-drivenlaw legal activity performed by humans by means of sources of norms such as statute and case law legal norms or policies that have been articulated in computer code automatic decision-making or predictions used in support derived from statistical/inductive methods statistical closure logical closure legal certainty + justice (proportionality, distribution, …) ? instrumentality (contextual policy goals) ?
  • 13.
    code-driven law data-driven law text-guidedlaw legal activity performed by humans by means of sources of norms such as statute and case law decision support systems statistical closure logical closure legal certainty + justice (proportionality, distribution, …) ? instrumentality (contextual policy goals) ? in so far humans are involved data-guided law automatic decision-making derived from statistical/inductive methods legal norms or policies that have been articulated in computer code
  • 14.
    code-driven law data-driven law text-guidedlaw legal activity performed by humans by means of sources of norms such as statute and case law decision support systems statistical closure logical closure legal certainty + justice (proportionality, distribution, …) ? instrumentality (contextual policy goals) ? data-guided law automatic decision-making derived from statistical/inductive methods instance + label instance + label instance + label instance + label ….. TRAINING DATASET legal norms or policies that have been articulated in computer code
  • 15.
    code-driven law data-driven law text-guidedlaw legal activity performed by humans by means of sources of norms such as statute and case law decision support systems statistical closure logical closure legal certainty + justice (proportionality, distribution, …) ? instrumentality (contextual policy goals) ? data-guided law automatic decision-making derived from statistical/inductive methods instance + label instance + label instance + label instance + label ….. is the training data set representative of the domain in which it will be applied? is the rationality behind labeling still the same? TRAINING DATASET legal norms or policies that have been articulated in computer code
  • 16.
    code-driven law data-driven law text-guidedlaw legal activity performed by humans by means of sources of norms such as statute and case law decision support systems statistical closure logical closure legal certainty + justice (proportionality, distribution, …) ? instrumentality (contextual policy goals) ? data-guided law automatic decision-making derived from statistical/inductive methods instance + label instance + label instance + label instance + label ….. is the training data set representative of the domain in which it will be applied? is the rationality behind labeling still the same? TRAINING DATASET is the training method producing adequate, reproducible, robust results? TRAINING METHOD legal norms or policies that have been articulated in computer code
  • 17.
    code-driven law data-driven law text-guidedlaw legal activity performed by humans by means of sources of norms such as statute and case law decision support systems statistical closure logical closure legal certainty + justice (proportionality, distribution, …) ? instrumentality (contextual policy goals) ? data-guided law automatic decision-making derived from statistical/inductive methods instance + label instance + label instance + label instance + label ….. is the training data set representative of the domain in which it will be applied? is the rationality behind labeling still the same? TRAINING DATASET is the training method producing adequate, reproducible, robust results? TRAINING METHOD → it can be contested legal norms or policies that have been articulated in computer code
  • 18.
    code-driven law data-driven law text-guidedlaw legal activity performed by humans by means of sources of norms such as statute and case law decision support systems statistical closure logical closure legal certainty + justice (proportionality, distribution, …) ? instrumentality (contextual policy goals) ? data-guided law automatic decision-making derived from statistical/inductive methods fact fact …. STRUCTURAL rule rule ….. CONTINGENT legal norms or policies that have been articulated in computer code
  • 19.
    code-driven law data-driven law text-guidedlaw legal activity performed by humans by means of sources of norms such as statute and case law decision support systems statistical closure logical closure legal certainty + justice (proportionality, distribution, …) ? instrumentality (contextual policy goals) ? data-guided law automatic decision-making derived from statistical/inductive methods fact fact …. STRUCTURAL rule rule ….. CONTINGENT were all the relevant rules considered? were all the relevant facts given? is the formalization correct? is the reasoning tool robust? legal norms or policies that have been articulated in computer code
  • 20.
    code-driven law data-driven law text-guidedlaw legal activity performed by humans by means of sources of norms such as statute and case law decision support systems statistical closure logical closure legal certainty + justice (proportionality, distribution, …) ? instrumentality (contextual policy goals) ? data-guided law automatic decision-making derived from statistical/inductive methods fact fact …. STRUCTURAL rule rule ….. CONTINGENT → it can be contested were all the relevant rules considered? were all the relevant facts given? is the formalization correct? is the reasoning tool robust? legal norms or policies that have been articulated in computer code
  • 21.
    code-driven law data-driven law text-guidedlaw legal activity performed by humans by means of sources of norms such as statute and case law decision support systems statistical closure logical closure legal certainty ? justice (proportionality, distribution, …) ? instrumentality (contextual policy goals) ? data-guided law automatic decision-making derived from statistical/inductive methods the COMPUTATIONAL LEGALISM “ideal” works only in so far we forget all what may go wrong legal norms or policies that have been articulated in computer code
  • 22.
    code-driven law data-driven law text-guidedlaw legal activity performed by humans by means of sources of norms such as statute and case law decision support systems statistical closure logical closure legal certainty ? justice (proportionality, distribution, …) ? instrumentality (contextual policy goals) ? data-guided law automatic decision-making derived from statistical/inductive methods is this all? legal norms or policies that have been articulated in computer code
  • 23.
    code-driven law data-driven law text-guidedlaw legal activity performed by humans by means of sources of norms such as statute and case law decision support systems statistical closure logical closure legal certainty ? justice (proportionality, distribution, …) + instrumentality (contextual policy goals) ? data-guided law automatic decision-making derived from statistical/inductive methods these principles could be to some extent formalized as meta-rules legal norms or policies that have been articulated in computer code
  • 24.
    code-driven law data-driven law text-guidedlaw legal activity performed by humans by means of sources of norms such as statute and case law decision support systems statistical closure logical closure legal certainty ? justice (proportionality, distribution, …) + instrumentality (contextual policy goals) + data-guided law automatic decision-making derived from statistical/inductive methods these principles could be to some extent formalized as meta-rules these goals could be made explicit legal norms or policies that have been articulated in computer code
  • 25.
    code-driven law data-driven law text-guidedlaw legal activity performed by humans by means of sources of norms such as statute and case law decision support systems statistical closure logical closure legal certainty ? justice (proportionality, distribution, …) + instrumentality (contextual policy goals) + data-guided law automatic decision-making derived from statistical/inductive methods these principles could be to some extent formalized as meta-rules these goals could be made explicit can we be sure about a certain formalization? legal norms or policies that have been articulated in computer code
  • 26.
    code-driven law data-driven law text-guidedlaw legal activity performed by humans by means of sources of norms such as statute and case law decision support systems statistical closure logical closure legal certainty ? justice (proportionality, distribution, …) ? instrumentality (contextual policy goals) + data-guided law automatic decision-making derived from statistical/inductive methods these principles could be to some extent formalized as meta-rules these goals could be made explicit never. can we be sure about a certain formalization? (cf. discussions on universal human rights, ethical frameworks) legal norms or policies that have been articulated in computer code
  • 27.
    code-driven law data-driven law text-guidedlaw legal activity performed by humans by means of sources of norms such as statute and case law decision support systems statistical closure logical closure legal certainty ? justice (proportionality, distribution, …) ? instrumentality (contextual policy goals) + data-guided law automatic decision-making derived from statistical/inductive methods these principles could be to some extent formalized as meta-rules these goals could be made explicit may we be ever ready for that? (explicit articulation of policy-maker purposes) never. can we be sure about a certain formalization? (cf. discussions on universal human rights, ethical frameworks) legal norms or policies that have been articulated in computer code
  • 28.
    code-driven law data-driven law text-guidedlaw legal activity performed by humans by means of sources of norms such as statute and case law decision support systems statistical closure logical closure legal certainty ? justice (proportionality, distribution, …) ? instrumentality (contextual policy goals) ? data-guided law automatic decision-making derived from statistical/inductive methods these principles could be to some extent formalized as meta-rules these goals could be made explicit may we be ever ready for that? (explicit articulation of policy-maker purposes) not to a full extent. never. can we be sure about a certain formalization? (cf. discussions on universal human rights, ethical frameworks) (it may require to be transparent on highly contestable issues) legal norms or policies that have been articulated in computer code
  • 29.
    code-driven law data-driven law text-guidedlaw legal activity performed by humans by means of sources of norms such as statute and case law decision support systems statistical closure logical closure legal certainty ? justice (proportionality, distribution, …) ? instrumentality (contextual policy goals) ? data-guided law automatic decision-making derived from statistical/inductive methods these principles could be to some extent formalized as meta-rules these goals could be made explicit may we be ever ready for that? (explicit articulation of policy-maker purposes) not to a full extent. never. can we be sure about a certain formalization? (cf. discussions on universal human rights, ethical frameworks) (it may require to be transparent on highly contestable issues) the COMPUTATIONAL JUSTICE “ideal” assumes perfect “conceptualization” legal norms or policies that have been articulated in computer code
  • 30.
    code-driven law data-driven law text-guidedlaw legal activity performed by humans by means of sources of norms such as statute and case law decision support systems legal certainty ? justice (proportionality, distribution) ? instrumentality (contextual policy goals) ? data-guided law automatic decision-making derived from statistical/inductive methods the COMPUTATIONAL JUSTICE “ideal” assumes perfect “conceptualization” the COMPUTATIONAL LEGALISM “ideal” works only in so far we forget all what may go wrong legal norms or policies that have been articulated in computer code
  • 31.
    code-driven law data-driven law text-guidedlaw legal activity performed by humans by means of sources of norms such as statute and case law decision support systems legal certainty ? justice (proportionality, distribution) ? instrumentality (contextual policy goals) ? data-guided law automatic decision-making derived from statistical/inductive methods the COMPUTATIONAL JUSTICE “ideal” assumes perfect “conceptualization” the COMPUTATIONAL LEGALISM “ideal” works only in so far we forget all what may go wrong let’s accept things may go wrong at all level legal norms or policies that have been articulated in computer code
  • 32.
    code-driven law data-driven law text-guidedlaw legal activity performed by humans by means of sources of norms such as statute and case law decision support systems legal certainty ? justice (proportionality, distribution) ? instrumentality (contextual policy goals) ? data-guided law automatic decision-making derived from statistical/inductive methods the COMPUTATIONAL JUSTICE “ideal” assumes perfect “conceptualization” the COMPUTATIONAL LEGALISM “ideal” works only in so far we forget all what may go wrong let’s accept things may go wrong at all level people or other actors should be able to appeal legal norms or policies that have been articulated in computer code
  • 33.
    code-driven law data-driven law text-guidedlaw legal activity performed by humans by means of sources of norms such as statute and case law decision support systems legal certainty ? justice (proportionality, distribution) ? instrumentality (contextual policy goals) ? data-guided law automatic decision-making derived from statistical/inductive methods the COMPUTATIONAL JUSTICE “ideal” assumes perfect “conceptualization” the COMPUTATIONAL LEGALISM “ideal” works only in so far we forget all what may go wrong let’s accept things may go wrong at all level let’s accept we need to integrate systematically feedbacks from higher-order courts, jurisprudence or other normative sources people or other actors should be able to appeal legal norms or policies that have been articulated in computer code
  • 34.
    code-driven law data-driven law text-guidedlaw legal activity performed by humans by means of sources of norms such as statute and case law decision support systems legal certainty ? justice (proportionality, distribution) ? instrumentality (contextual policy goals) ? data-guided law automatic decision-making derived from statistical/inductive methods let’s accept things may go wrong at all level let’s accept we need to integrate systematically feedbacks from higher-order courts, jurisprudence or other normative sources people or other actors should be able to appeal courts or other actors should be able to quash legal norms or policies that have been articulated in computer code the COMPUTATIONAL JUSTICE “ideal” assumes perfect “conceptualization” the COMPUTATIONAL LEGALISM “ideal” works only in so far we forget all what may go wrong
  • 35.
    including computational actors possibilityof continuous, automated testing/verification code-driven law data-driven law text-guided law legal activity performed by humans by means of sources of norms such as statute and case law decision support systems legal certainty ? justice (proportionality, distribution) ? instrumentality (contextual policy goals) ? data-guided law automatic decision-making derived from statistical/inductive methods let’s accept things may go wrong at all level let’s accept we need to integrate systematically feedbacks from higher-order courts, jurisprudence or other normative sources people or other actors should be able to appeal courts or other actors should be able to quash legal norms or policies that have been articulated in computer code the COMPUTATIONAL JUSTICE “ideal” assumes perfect “conceptualization” the COMPUTATIONAL LEGALISM “ideal” works only in so far we forget all what may go wrong
  • 36.
    including computational actors possibilityof continuous, automated testing/verification code-driven law data-driven law text-guided law derived from legal norms or policies that have been articulated in computer code (or from other methods) lower authority w.r.t. human authorities!!! legal activity performed by humans by means of sources of norms such as statute and case law decision support systems legal certainty ? justice (proportionality, distribution) ? instrumentality (contextual policy goals) ? data-guided law automatic decision-making derived from statistical/inductive methods the COMPUTATIONAL JUSTICE “ideal” assumes perfect “conceptualization” the COMPUTATIONAL LEGALISM “ideal” works only in so far we forget all what may go wrong let’s accept things may go wrong at all level let’s accept we need to integrate systematically feedbacks from higher-order courts, jurisprudence or other normative sources people or other actors should be able to appeal courts or other actors should be able to quash code-guided law monolithical systems ecological system including interfaces with humans
  • 37.
    including computational actors possibilityof continuous, automated testing/verification code-driven law data-driven law text-guided law derived from legal norms or policies that have been articulated in computer code (or from other methods) lower authority w.r.t. human authorities!!! legal activity performed by humans by means of sources of norms such as statute and case law decision support systems legal certainty ? justice (proportionality, distribution) ? instrumentality (contextual policy goals) ? data-guided law automatic decision-making derived from statistical/inductive methods the COMPUTATIONAL JUSTICE “ideal” assumes perfect “conceptualization” the COMPUTATIONAL LEGALISM “ideal” works only in so far we forget all what may go wrong let’s accept things may go wrong at all level let’s accept we need to integrate systematically feedbacks from higher-order courts, jurisprudence or other normative sources people or other actors should be able to appeal courts or other actors should be able to quash code-guided law monolithical systems ecological system including interfaces with humans a serious technological gap exists today...
  • 38.
  • 39.
    A tentative ontology SOFTWARE HARDWARE ●physical device ● when running ⇒ physical process ● situated in a physical environment ● symbolic device ● when running ⇒ symbolic process ● relies on physical processes
  • 40.
    ? A tentative ontology SOFTWARE HARDWARENORMWARE ● physical device ● when running ⇒ physical process ● situated in a physical environment ● symbolic device ● when running ⇒ symbolic process ● relies on physical processes ● ? ● when running ⇒ ? ● relies on ?
  • 41.
    ? A tentative ontology SOFTWARE HARDWARENORMWARE ● physical device ● when running ⇒ physical process ● situated in a physical environment ● symbolic device ● when running ⇒ symbolic process ● relies on physical processes ● ? ● when running ⇒ ? ● relies on ? ARTIFACT dimension PROCESS dimension
  • 42.
    Normware as artifacts: directivesconcerning regulation the cookie jar must be full you are prohibited to eat cookies you can not eat cookies
  • 43.
    Normware as artifacts: directivesconcerning regulation aiming to regulate behaviour aiming to regulate situations in the world the cookie jar must be full you are prohibited to eat cookies you can not eat cookies
  • 44.
    we can haveexplicit normative specifications… Normware as artifacts: directives concerning regulation access-control policies languages based on deontic logic languages based on Hohfeld’s primitives
  • 45.
    but programs inthemselves are already mandatory in nature! a := 2 + 2 system has to perform 2 + 2… ?mother(maggie, bart) system has to prove that… animal :- dog. system has to make animal true if dog is true Normware as artifacts: directives concerning regulation
  • 46.
    Normware as artifacts: directivesconcerning terminology the cookie jar must be full you are prohibited to eat cookies you can not eat cookies what is a cookie? what does it mean to be full? what is eating? who is you? what is a jar?
  • 47.
    Normware as artifacts: directivesconcerning expectations the cookie jar must be full you are prohibited to eat cookies you can not eat cookies eating cookies → cookies are destroyed → the jar is not full practical normative reasoning always require some world knowledge
  • 48.
    Normware as artifacts: devicesintended to regulate doors regulate entrances semaphores regulate traffic
  • 49.
    Normware as artifacts: devicesintended to regulate doors regulate entrances semaphores regulate traffic we do not have access to the inner decision-making mechanism that brings to the current output state!
  • 50.
    Normware as artifacts: devicesintended to regulate black-boxes (eg. ML models) are also artifacts expressing some form of normativity/normality “is this a cat?” “how to (best) behave in certain conditions?”
  • 51.
    Normware as artifacts: devicesintended to regulate black-boxes (eg. ML models) are also artifacts expressing some form of normativity/normality “is this a cat?” “how to (best) behave in certain conditions?” from a functional point of view, they also count as normware!
  • 52.
    Whether artificial ornatural, designed or emergent, what counts in control is ● the existence of some reference (the target of control), ● which the entity is set to either approach or avoid (the direction of control). Normware as processes: regulation as control
  • 53.
    Whether artificial ornatural, designed or emergent, what counts in control is ● the existence of some reference (the target of control), ● which the entity is set to either approach or avoid (the direction of control). behavioural or situational space intervention current state target state Normware as processes: regulation as control
  • 54.
    what defines thereferences though? target state current state direction Normware as processes: regulation as control
  • 55.
    Whether artificial ornatural, designed or emergent, what counts in control is ● the existence of some reference (the target of control), ● which the entity is set to either approach or avoid (the direction of control). by defining directives by this control signature (target, direction), any regulative mechanisms can be abstracted from its implementation. direction Normware as processes: regulation as control current state target state
  • 56.
    indeterminacy of references Normwareas processes: higher-order indetermination
  • 57.
    indeterminacy of references indeterminacyof directives Normware as processes: higher-order indetermination
  • 58.
    indeterminacy of referencesantinomies indeterminacy of directives Normware as processes: higher-order indetermination
  • 59.
    indeterminacy of referencesantinomies indeterminacy of directives Normware as processes: higher-order indetermination mechanisms of conflict resolution are needed at systematic level!
  • 60.
  • 61.
    Normware: first-order control GOALACTION labelled dataset ML black box instruction operation data-driven code-driven
  • 62.
    Normware: first-order control GOALACTION labelled dataset ML black box instruction operation data-driven code-driven pool of actions In both cases, some method has been used to select the action from a pool of actions…
  • 63.
    GOAL ACTION But then,where the goal comes from? pool of actions Normware: first-order control labelled dataset ML black box instruction operation data-driven code-driven
  • 64.
    TACTICAL GOAL ACTION poolof actions Normware: second-order control STRATEGIC GOAL Adding depth!
  • 65.
    pool of goals TACTICALGOAL ACTION pool of actions Normware: second-order control STRATEGIC GOAL Adding depth!
  • 66.
    pool of goals TACTICALGOAL ACTION pool of actions Normware: second-order control STRATEGIC GOAL Sileno, G., Boer, A. and van Engers, T., The Role of Normware in Trustworthy and Explainable AI, Proceedings of XAILA workshop: Explainable AI and Law, in conjunction with JURIX 2018 trustworthy AI and explainable AI issues in ML due to lack of the strategic component Adding depth!
  • 67.
    TACTICAL GOAL ACTION Normware:second-order control STRATEGIC GOAL cybernetic view on systems: policy, intelligence, operations
  • 68.
    TACTICAL GOAL ACTION Normware:second-order control STRATEGIC GOAL …yet it is about a single “organism”, not an “ecology” “TOTALITARIAN” architecture cybernetic view on systems: policy, intelligence, operations
  • 69.
    Normware: plural second-ordercontrol we need to acknowledge the presence of several autonomous entities,
  • 70.
    Normware: plural second-ordercontrol we need to acknowledge the presence of several autonomous entities,
  • 71.
    Normware: plural second-ordercontrol we need to acknowledge the presence of several autonomous entities, and adequate conflict resolution mechanisms
  • 72.
    Normware: plural second-ordercontrol resolution mechanisms may also result from a similar process
  • 73.
    Normware: plural second-ordercontrol resolution mechanisms may also result from a similar process “PARTICIPATORY” architecture
  • 74.
    Normware: plural second-ordercontrol resolution mechanisms may also result from a similar process “PARTICIPATORY” architecture Sileno, G., Grosso, P., Accounting Value Effects for Responsible Networking. Proceedings of ACM SIGCOMM 2021 Workshop on Technologies, Applications, and Uses of a Responsible Internet (TAURIN2021).
  • 75.
    ? SOFTWARE HARDWARE NORMWARE physical device whenrunning ⇒ physical process situated in a physical environment symbolic device when running ⇒ symbolic process relies on physical processes coordination device when running ⇒ coordination process relies on symbolic (possibly hard-coded) processes A less tentative ontology
  • 76.
    Key points ● Thereis continuity between institutional and computational activities. ● We don’t need institutions to become more mechanical, institutional constructs need to be brought into the computational realm. ● As we have a plurality of normative sources, we need a plurality of computational normative sources. normware
  • 77.
    mechanizing law introducing legitimatenormative processes within the computational realm SI! NO Contemporary socio-technical challenge:
  • 78.
    Contemporary socio-technical challenge: mechanizinglaw introducing legitimate normative processes within the computational realm SI! code-driven law data-driven law normware law-makers, policy-makers, citizens, … NO
  • 79.
    Code-driven Law NO, NormwareSI! 3 November 2022, CRCL conference “Computational Law on Edge”, Brussels Giovanni Sileno, g.sileno@uva.nl Socially Intelligent Artificial Systems (SIAS), Informatics Institute, University of Amsterdam