The document discusses defining open source artificial intelligence (AI). It outlines the Open Source Initiative's (OSI) mission to educate about and advocate for open source software. The OSI is working to define open source AI to clarify expectations and match them with policy goals of transparency, trustworthiness, etc. A proposed definition grants users four freedoms: to study and inspect AI systems, use them without permission, modify them, and share modified versions. The OSI will hold an in-person meeting in early 2024 to further develop an open source AI definition.

1. TOWARDS A DEFINITION
OPEN SOURCE AI
Stefano Maffulli - Executive director
Open Source Initiative
OpenChain webinar
Nov 27, 2023

2. Mission
The Open Source Initiative (OSI) is the steward of the Open Source
Definition, the foundation for the multi-billion dollar open source
software ecosystem.
A non-profit corporation with global scope formed to educate about
and advocate for the benefits of open source and to build bridges
among different constituencies in the open source community.

3. License and Legal
An anchor for open community
consensus on what constitutes
Open Source. We protect the
Open Source principles,
enforcing the marks “Certified
Open Source” and “Open
Source Approved License”.
Policy & Standards
Monitoring policy and
standards setting
organizations. Supporting
legislators and policy
makers educating them
about the Open Source
ecosystem, its role in
innovation and its value
for an open future.
Advocacy & Outreach
Leading global conversations
with non-profits, developers
and lawyers to improve the
understanding of Open
Source principles and
practice. OSI investigates the
impacts of ongoing debates
around Open Source, from
artificial intelligence to
security.
We define Open Source
Simplifying software development and procurement
Our programs

4. Our projects
License and Legal Policy & Standards Advocacy & Outreach
OpenSource.net
Deep Dive: Defining
Open Source AI

5. The Cyber Resilience Act
Final weeks of negotiations

6. TECHNOLOGY EVOLVES
Distribution and execution of software change
And Open Source must adapt.
1
Deep Dive: Defining Open Source AI

7. WHY NOW
The EU AI Act
A provider of a foundation
model shall, prior to making it
available on the market or
putting it into service, ensure
that it is compliant with the
requirements set out in this
Article, regardless of whether it
is provided as a standalone
model or embedded in an AI
system or a product, or
provided under free and open
source licences, as a service,
as well as other distribution
channels.
7
The confusion on the market
And the emergence of
gatekeepers
The speed of evolution
Sooo many models released so
quickly, so many new concerns
raised each time by so many
different set of people…
The lack of clarity favors those
who scream the loudest, not
the most competent or
reasonable.

8. Data > training > weights > model > app
Data
Web scraping, git
repositories, wikipedia,
internet archive… lots
of data!
Input
Training
Clean up the data, add
tags, classify
Architecture
Weights
Tuning the model to
the task.
Learning
Model
Train and test the
model against a
testing dataset. Wrong
answers? Rinse and
repeat.
Test
Application
Use the model in
applications.
Application

9. Do models have copyright?
Maybe.
Data may have copyright
Datasets may have more than copyright
9

10. This image has no copyright
Or does it?

11. The core issues of AI
The data
The models
The knowledge
The hardware
The uses
The social norms
11

12. The objective for 2024
Defining Open Source AI

13. How Open Source came to be
1: Legal framework
Copyright applied to
software, first.
This new artifact
became privatized
work.
Researchers
complained.
2: Principles
The GNU Manifesto
lays the ground to
oppose privatization.
A community forms
around these
principles.
3: Licenses
Copyleft is a hack on
copyright.
Incorporating the
principles, serving as
the Constitution of a
forming community.
13

14. Restrictions are an obstacle to innovation
[…] much wasteful duplication of system programming effort will be
avoided
“If I like a program, I must share it with other people who like it.”
Golden rule, GNU Manifesto
14
Schools will be able to provide a much more educational environment… by
encouraging all students to study and improve the system code
[…] the overhead of considering who owns the system software and what
one is or is not entitled to do with it will be lifted

15. 15
What is Open Source (also known as free software)
A program is free software if the program's users have the four
essential freedoms:
0. The freedom to run the program as you wish, for any purpose
1. The freedom to study how the program works, and change it so
it does your computing as you wish. Access to the source code is
a precondition for this.
2. The freedom to redistribute copies so you can help others.
3. The freedom to distribute copies of your modified versions to
others (freedom 3). By doing this you can give the whole
community a chance to benefit from your changes. Access to the
source code is a precondition for this.
●

16. 16
What is Open Source
Open Source gives users and developers the ability to
● decide for themselves how and where to use the
technology
● without the need to engage with another party

17. Golden Rule applied to AI
If I like an AI system I must share it with other people who like it.

18. What we’ve learned so far
● We need to define Open Source AI, in general, not just
machine learning
● OECD’s definition of AI is well accepted (caveat: decisions)
“An AI system is a machine-based system that, for explicit or
implicit objectives, infers, from the input it receives, how to
generate outputs such as predictions, content,
recommendations, or decisions that can influence physical or
virtual environments. Different AI systems vary in their levels of
autonomy and adaptiveness after deployment.” (2023)

19. Matching expectations
AI deserves to enjoy the
benefits of Open Source
● autonomy
● transparency
● collaborative
improvement
Policy makers, academia and
industry are focusing on
● transparency
● trustworthiness
● reliability
● transparency
● explainability
● fairness
● safety etc

20. It should be possible for Open Source AI to be “good”
● trustworthy
● transparent
● safe
● reliable
● fair
● responsible
● explainable
● …
● etc.
“good”
Not all Open Source AI will
be good
But all Open Source AI can
be good

21. What basic freedoms do we need?

22. Definition of AI system
Preamble
Out of scope issues
4 freedoms
License checklist

23. What is Open Source AI
To be Open Source, an AI system needs to make its
components available under licenses that individually
grant [users of the system] the freedoms to:
● Study how the system works and inspect its
components.
● Use the system for any purpose and without having to
ask for permission.
● Modify the system to change its outputs to adapt to
your needs.
● Share the system with or without modifications, for
any purpose.

24. Comment draft v. 0.0.3 of the Open Source AI Definition
24
https://hackmd.io/@opensourceinitiative/osaid-0-0-3

25. Next steps
In-person meeting in Brussels, ~Feb 1-4, 2024
Working on a full calendar for 2024

26. Thanks!
Become a member of OSI
@ed@opensource.org
stefano@opensource.org
26
https://members.opensource.org/join
Support more workshops in 2024