SlideShare a Scribd company logo
1 of 27
Ethics of AI
An engineering perspective.
Francisco Curado Teixeira
(francisco.curado.teixeira@gmail.com)
ECA/UN Youth Innovation Bootcamp on Emerging Technologies
23 Feb 2021
Acknowledgments: Project RETIOT, DETI-IEETA, Univ. of Aveiro
(http://wiki.ieeta.pt/wiki/index.php/Institute_of_Electronics_and_Informatics_Engineering_of_Aveiro)
Given that AI is a powerful technology, we have a
moral obligation to use it well, to promote the positive
aspects and avoid or mitigate the negative ones.
Russell, S. J. & Norvig, P., 2020. Artificial Intelligence - A Modern Approach. Pearson
-----------------
If the feedback is built into a machine that cannot be inspected until the final
goal is attained, the possibilities for catastrophe are greatly increased. I should
very much hate to ride on the first trial of an automobile regulated by
photoelectric feedback devices, unless there were somewhere a handle by
which I could take over control if I found myself driving smack into a tree.
Norbert Wiener, 1964. God & Golem, Inc.
AI Ethics by Design - a Case Study
Problem Motivation.
AI Ethics by Design - a Case Study
Detection & Tracking of moving subjects in indoor /
outdoor spaces.
Applications (Public Surveillance & Security not contemplated)
Hospitals and Healthcare
● Monitor space occupation and movement of people
inside the building
● Detect and Alert for falls and inanimate people
City traffic management
● Traffic assessment & route planning
● Safety of pedestrians & bicyclists: warn for imminent
run-overs and collisions
See Credits list: [A]
See Credits list: [B]
See Credits list: [D]
AI Ethics by Design - a Case Study
Ethical Considerations.
AI Ethics by Design - a Case Study
Detection & Tracking of moving subjects in indoor /
outdoor spaces.
Ethical and operational requirements
Inclusiveness
● Account for diversity of subjects and situations
○ Humans, animals, vehicles, robots...
○ Apply to normal daily activities and detect abnormal behaviour
○ Account for people with disabilities
Privacy compliance
● Avoid acquiring data that permits personal identification
Safety
● Technology must be safe for humans and animals
AI Ethics by Design - a Case Study
Detection & Tracking of moving subjects in indoor / outdoor spaces.
Conventional solution: cameras
Advantages
● Well-established technology
● Low-cost H/W & CV software tools
Problems
● Privacy violation
● Illegal in most countries See Credits list: [C]
AI Ethics by Design - a Case Study
Technical Solutions.
Detection & Tracking of moving subjects in indoor / outdoor spaces.
Nonintrusive solutions: RADAR
Advantages
● Nonintrusive: no ID information in radar data
● Automotive Radar: compact, affordable, versatile FMCW tech.
● Safe for humans: 77-81 GHz
● Robust against environmental disturbances
● Appearing in mobile phones
Limitations
● No direct ‘picture’ acquisition
● Feature extraction highly dependent on radar hardware
AI Ethics by Design - a Case Study
Detection & Tracking of moving subjects in
indoor / outdoor spaces.
Nonintrusive solutions: RADAR
Radar working Principles and Techniques
● Cloud point of reflecting objects
● Rich data: range, angle, velocity,
intensity...
Problems
● Noisy raw data: clutter, multi-path...
● Complex feature extraction
AI Ethics by Design - Case Study
Detection & Tracking of moving subjects in
indoor / outdoor spaces.
Nonintrusive solutions: RADAR
Radar working Principles and Techniques
● Cloud point of reflecting objects
● Rich data: range, angle, velocity,
intensity...
Problems
● Noisy raw data: clutter, multi-path...
● Complex feature extraction
AI Ethics by Design - Case Study
Detection & Tracking of moving subjects in
indoor / outdoor spaces.
Nonintrusive solutions: RADAR
Radar working Principles and Techniques
● Cloud point of reflecting objects
● Rich data: range, velocity, intensity...
Problems
● Noisy raw data: clutter, multi-path...
● Complex feature extraction
AI Ethics by Design - Case Study
Detection & Tracking of moving subjects in
indoor / outdoor spaces.
Nonintrusive solutions: RADAR
Motion detection Principles and Techniques
● Point cloud data: range, angle, Doppler,
Intensity…
● Cluster points: range & Doppler (velocity)
● Use Doppler patterns for classification
● Complementary: signal intensity (RCS)
AI Ethics by Design - Case Study
Radar geometry in 2D.
Source: [4]
Illustration of a person walking and a robot moving
and the respective point clouds. Source: [5]
Detection & Tracking of moving subjects in
indoor / outdoor spaces.
Nonintrusive solutions: RADAR
Complexity
● Advanced signal processing (raw data):
ADC, 3D FFT, Clustering, Data-
Association, Tracking…
● Novel feature extraction techniques
Advantages
● Diversity of available feature types (e.g.: RCS)
and dimensionality (e.g. multi-frame)
AI Ethics by Design - Case Study
Doppler pattern of mannequin or robot. Source: [5]
Doppler pattern of person walking. Source: [5]
AI Ethics by Design - a Case Study
Detection & Tracking of moving subjects in indoor / outdoor spaces.
Alternative solution: Low-resolution Thermal camera
Advantages
● Nonintrusive: no personal ID information
● Well-established, affordable technology
● Compact
Limitations
● Sensitivity to environmental disturbances
● Controversial in public spaces: similarity to Video & 3D cameras
AI Ethics by Design - a Case Study
Adopted Solution.
AI Ethics by Design - a Case Study
Anonymized Detection & Tracking of moving subjects in indoor / outdoor
spaces.
Preferred solution: RADAR-only sensing
Advantages of Radar Data
● Subject classification based on velocity (Doppler) patterns
● Multi-class classification: pedestrians, bicyclists, cars, animals...
Problems:
● Limited availability of AI tools and scarcity of labeled data-sets
● Huge amount of data supplied by the radar
● The majority of radar frames consist of noise and requires human inspection
● Radar frames are NOT easily interpretation by humans
● Data labeling is a highly time-consuming, expensive and tedious process
??
*!* ?!
AI Ethics by Design - a Case Study
Anonymized Detection & Tracking of moving
subjects in indoor / outdoor spaces.
Proposed: Training with Radar + Thermal data
Data acquisition and Labeling: Initial setup
● Radar board + Raspberry Pi CPU + camera
● WiFi + PC
● Using camera images to select radar frames allows
for automatic or supervised labeling of subjects
Final setup
● Low-res, low-cost thermal camera replacing RPi
camera
AI Ethics by Design - a Case Study
Anonymized Detection & Tracking of moving
subjects in indoor / outdoor spaces.
Proposed: Training with Radar + Thermal data
Deployment in Laboratorial / Academic environment
● Nonintrusive: no personal ID information
● Inform the community about the Experiment
Advantages
● Favorable context for system deployment
● Radar and Camera can be left shooting without
supervision
AI Ethics by Design - a Case Study
Results.
Anonymized Detection & Tracking of
moving subjects in indoor / outdoor spaces.
Model Learning: RADAR-only deployed on a
mobile robot
Multi-class Classification
● People
● Mannequin
● Person in wheelchair
● Robots
AI Ethics by Design - a Case Study
Anonymized Detection & Tracking of
moving subjects in indoor / outdoor spaces.
Model Learning: RADAR-only deployed on a
mobile robot
Results with initial datasets (25431 samples)
after hyper-parameter optimization
● Accuracy: Artificial Neural Network (ANN)
○ Doppler only: 92%
○ Doppler + RCS: 96%
● Accuracy: Random Forest (RF)
○ Doppler only: 92%
○ Doppler + RCS: 95%
AI Ethics by Design - a Case Study
Anonymized Detection & Tracking of
moving subjects in indoor / outdoor spaces.
Model Learning: RADAR-only sensing
Increasing the dataset
● Start with a small initial labeled dataset and
trained model (classifier)
● Apply Active and Cooperative Learning
to enlarge the labeled dataset and improve
the model
● Use synthetic or augmented data: physical
models; stochastic data-augmentation
AI Ethics by Design - a Case Study
Active
Learning
loop
AI Ethics by Design - a Case Study
Conclusions.
Anonymized Detection & Tracking of moving subjects in indoor / outdoor
spaces.
Anonymized detection, classification, and tracking
● It is possible to implement this type of system without using vision
● It is crucial to address ethical issues in the design phase
● Ethical considerations may condition the choice of H/W and S/W Tools
● Alternative technologies and methods may present increased complexity
and cost but may reveal more robust and versatile
AI Ethics by Design - a Case Study
Thank you
Acknowledgments: Project RETIOT, DETI-IEETA, Univ. of Aveiro
(http://wiki.ieeta.pt/wiki/index.php/Institute_of_Electronics_and_Informatics_Engineering_of_Aveiro)
References
1. Russell, S. J. & Norvig, P., 2020. Artificial Intelligence. - A Modern Approach. Pearson
2. Dubber, M. D.; Pasquale, F. & Das, S. (Eds.), 2020. The Oxford Handbook of ETHICS OF AI. Oxford University
Press
3. Ng, Andrew, 2018. Machine Learning Yearning. www.dbooks.org
4. Livshitz, M. 2018. Tracking radar targets with multiple reflection points. https://e2e.ti.com/cfs-
file/__key/communityserver-discussions-components-files/1023/ [Online accessed 13-June-2019].
5. Castanheira, J., 2019. Machine Learning Methods for Radar-Based People Detection. Master thesis. Universidade
de Aveiro.
Figures’ credits
A. By Micov at English Wikipedia, CC BY-SA 3.0: https://commons.wikimedia.org/w/index.php?curid=47564675
B. Dllu: https://commons.wikimedia.org/wiki/User:Dllu
C. Paweł Zdziarski, CC BY-SA 3.0 via Wikimedia Commons <http://creativecommons.org/licenses/by-sa/3.0/>
D. US News - Health: https://health.usnews.com/health-news/patient-advice/slideshows/14-ways-to-protect-seniors-from-falls

More Related Content

Similar to Ethics of AI - An Engineering Perspective

Autonomous Vehicles: the Intersection of Robotics and Artificial Intelligence
Autonomous Vehicles: the Intersection of Robotics and Artificial IntelligenceAutonomous Vehicles: the Intersection of Robotics and Artificial Intelligence
Autonomous Vehicles: the Intersection of Robotics and Artificial IntelligenceWiley Jones
 
AUTOMATIC ATTENDANCE SYSTEM MANAGEMENT USING RASPBERRY PI WITH ULTRASONIC SENSOR
AUTOMATIC ATTENDANCE SYSTEM MANAGEMENT USING RASPBERRY PI WITH ULTRASONIC SENSORAUTOMATIC ATTENDANCE SYSTEM MANAGEMENT USING RASPBERRY PI WITH ULTRASONIC SENSOR
AUTOMATIC ATTENDANCE SYSTEM MANAGEMENT USING RASPBERRY PI WITH ULTRASONIC SENSORIRJET Journal
 
[VFS 2019] Human Activity Recognition Approaches
[VFS 2019] Human Activity Recognition Approaches [VFS 2019] Human Activity Recognition Approaches
[VFS 2019] Human Activity Recognition Approaches Nexus FrontierTech
 
IRJET- Object Detection and Recognition for Blind Assistance
IRJET- Object Detection and Recognition for Blind AssistanceIRJET- Object Detection and Recognition for Blind Assistance
IRJET- Object Detection and Recognition for Blind AssistanceIRJET Journal
 
From RGB led laboratory to servomotor control with websockets and IoT as educ...
From RGB led laboratory to servomotor control with websockets and IoT as educ...From RGB led laboratory to servomotor control with websockets and IoT as educ...
From RGB led laboratory to servomotor control with websockets and IoT as educ...German Carro Fernandez
 
Io t research_arpanpal_iem
Io t research_arpanpal_iemIo t research_arpanpal_iem
Io t research_arpanpal_iemArpan Pal
 
Emerging engineering issues for building large scale AI systems By Srinivas P...
Emerging engineering issues for building large scale AI systems By Srinivas P...Emerging engineering issues for building large scale AI systems By Srinivas P...
Emerging engineering issues for building large scale AI systems By Srinivas P...Analytics India Magazine
 
Arpan pal roboticsensing_sw2015
Arpan pal roboticsensing_sw2015Arpan pal roboticsensing_sw2015
Arpan pal roboticsensing_sw2015Arpan Pal
 
Pydata Chicago - work hard once
Pydata Chicago - work hard oncePydata Chicago - work hard once
Pydata Chicago - work hard onceJi Dong
 
Intelligent Embedded Systems (Robotics)
Intelligent Embedded Systems (Robotics)Intelligent Embedded Systems (Robotics)
Intelligent Embedded Systems (Robotics)Adeyemi Fowe
 
Innovation fund themed competition webinar - session 2
Innovation fund themed competition webinar - session 2Innovation fund themed competition webinar - session 2
Innovation fund themed competition webinar - session 2Heather-Fiona Egan
 
Ambient Intelligence: Definitions and Application Areas
Ambient Intelligence: Definitions and Application AreasAmbient Intelligence: Definitions and Application Areas
Ambient Intelligence: Definitions and Application AreasFulvio Corno
 
HOW TO WASTE YOUR TIME ON SIMPLE THINGS DONT JUST FEEL INSTEAD BLAME OTHERS A...
HOW TO WASTE YOUR TIME ON SIMPLE THINGS DONT JUST FEEL INSTEAD BLAME OTHERS A...HOW TO WASTE YOUR TIME ON SIMPLE THINGS DONT JUST FEEL INSTEAD BLAME OTHERS A...
HOW TO WASTE YOUR TIME ON SIMPLE THINGS DONT JUST FEEL INSTEAD BLAME OTHERS A...lanaw86385
 
Advance Intelligent Video Surveillance System Using OpenCV
Advance Intelligent Video Surveillance System Using OpenCVAdvance Intelligent Video Surveillance System Using OpenCV
Advance Intelligent Video Surveillance System Using OpenCVIRJET Journal
 
ILT202411111111111111111111111111111.pdf
ILT202411111111111111111111111111111.pdfILT202411111111111111111111111111111.pdf
ILT202411111111111111111111111111111.pdfw7823125
 
DSNet Joint Semantic Learning for Object Detection in Inclement Weather Condi...
DSNet Joint Semantic Learning for Object Detection in Inclement Weather Condi...DSNet Joint Semantic Learning for Object Detection in Inclement Weather Condi...
DSNet Joint Semantic Learning for Object Detection in Inclement Weather Condi...IRJET Journal
 
Sanjaya: A Blind Assistance System
Sanjaya: A Blind Assistance SystemSanjaya: A Blind Assistance System
Sanjaya: A Blind Assistance SystemIRJET Journal
 

Similar to Ethics of AI - An Engineering Perspective (20)

Autonomous Vehicles: the Intersection of Robotics and Artificial Intelligence
Autonomous Vehicles: the Intersection of Robotics and Artificial IntelligenceAutonomous Vehicles: the Intersection of Robotics and Artificial Intelligence
Autonomous Vehicles: the Intersection of Robotics and Artificial Intelligence
 
AUTOMATIC ATTENDANCE SYSTEM MANAGEMENT USING RASPBERRY PI WITH ULTRASONIC SENSOR
AUTOMATIC ATTENDANCE SYSTEM MANAGEMENT USING RASPBERRY PI WITH ULTRASONIC SENSORAUTOMATIC ATTENDANCE SYSTEM MANAGEMENT USING RASPBERRY PI WITH ULTRASONIC SENSOR
AUTOMATIC ATTENDANCE SYSTEM MANAGEMENT USING RASPBERRY PI WITH ULTRASONIC SENSOR
 
[VFS 2019] Human Activity Recognition Approaches
[VFS 2019] Human Activity Recognition Approaches [VFS 2019] Human Activity Recognition Approaches
[VFS 2019] Human Activity Recognition Approaches
 
IRJET- Object Detection and Recognition for Blind Assistance
IRJET- Object Detection and Recognition for Blind AssistanceIRJET- Object Detection and Recognition for Blind Assistance
IRJET- Object Detection and Recognition for Blind Assistance
 
From RGB led laboratory to servomotor control with websockets and IoT as educ...
From RGB led laboratory to servomotor control with websockets and IoT as educ...From RGB led laboratory to servomotor control with websockets and IoT as educ...
From RGB led laboratory to servomotor control with websockets and IoT as educ...
 
Io t research_arpanpal_iem
Io t research_arpanpal_iemIo t research_arpanpal_iem
Io t research_arpanpal_iem
 
Emerging engineering issues for building large scale AI systems By Srinivas P...
Emerging engineering issues for building large scale AI systems By Srinivas P...Emerging engineering issues for building large scale AI systems By Srinivas P...
Emerging engineering issues for building large scale AI systems By Srinivas P...
 
Ipen2018
Ipen2018Ipen2018
Ipen2018
 
Arpan pal roboticsensing_sw2015
Arpan pal roboticsensing_sw2015Arpan pal roboticsensing_sw2015
Arpan pal roboticsensing_sw2015
 
Pydata Chicago - work hard once
Pydata Chicago - work hard oncePydata Chicago - work hard once
Pydata Chicago - work hard once
 
Intelligent Embedded Systems (Robotics)
Intelligent Embedded Systems (Robotics)Intelligent Embedded Systems (Robotics)
Intelligent Embedded Systems (Robotics)
 
Innovation fund themed competition webinar - session 2
Innovation fund themed competition webinar - session 2Innovation fund themed competition webinar - session 2
Innovation fund themed competition webinar - session 2
 
Ambient Intelligence: Definitions and Application Areas
Ambient Intelligence: Definitions and Application AreasAmbient Intelligence: Definitions and Application Areas
Ambient Intelligence: Definitions and Application Areas
 
HOW TO WASTE YOUR TIME ON SIMPLE THINGS DONT JUST FEEL INSTEAD BLAME OTHERS A...
HOW TO WASTE YOUR TIME ON SIMPLE THINGS DONT JUST FEEL INSTEAD BLAME OTHERS A...HOW TO WASTE YOUR TIME ON SIMPLE THINGS DONT JUST FEEL INSTEAD BLAME OTHERS A...
HOW TO WASTE YOUR TIME ON SIMPLE THINGS DONT JUST FEEL INSTEAD BLAME OTHERS A...
 
Advance Intelligent Video Surveillance System Using OpenCV
Advance Intelligent Video Surveillance System Using OpenCVAdvance Intelligent Video Surveillance System Using OpenCV
Advance Intelligent Video Surveillance System Using OpenCV
 
ILT202411111111111111111111111111111.pdf
ILT202411111111111111111111111111111.pdfILT202411111111111111111111111111111.pdf
ILT202411111111111111111111111111111.pdf
 
Major PRC-1 ppt.pptx
Major PRC-1 ppt.pptxMajor PRC-1 ppt.pptx
Major PRC-1 ppt.pptx
 
DSNet Joint Semantic Learning for Object Detection in Inclement Weather Condi...
DSNet Joint Semantic Learning for Object Detection in Inclement Weather Condi...DSNet Joint Semantic Learning for Object Detection in Inclement Weather Condi...
DSNet Joint Semantic Learning for Object Detection in Inclement Weather Condi...
 
Sanjaya: A Blind Assistance System
Sanjaya: A Blind Assistance SystemSanjaya: A Blind Assistance System
Sanjaya: A Blind Assistance System
 
CV VD Mohire-Research
CV VD Mohire-ResearchCV VD Mohire-Research
CV VD Mohire-Research
 

Recently uploaded

Key Features Of Token Development (1).pptx
Key  Features Of Token  Development (1).pptxKey  Features Of Token  Development (1).pptx
Key Features Of Token Development (1).pptxLBM Solutions
 
Gen AI in Business - Global Trends Report 2024.pdf
Gen AI in Business - Global Trends Report 2024.pdfGen AI in Business - Global Trends Report 2024.pdf
Gen AI in Business - Global Trends Report 2024.pdfAddepto
 
Scanning the Internet for External Cloud Exposures via SSL Certs
Scanning the Internet for External Cloud Exposures via SSL CertsScanning the Internet for External Cloud Exposures via SSL Certs
Scanning the Internet for External Cloud Exposures via SSL CertsRizwan Syed
 
Unraveling Multimodality with Large Language Models.pdf
Unraveling Multimodality with Large Language Models.pdfUnraveling Multimodality with Large Language Models.pdf
Unraveling Multimodality with Large Language Models.pdfAlex Barbosa Coqueiro
 
Nell’iperspazio con Rocket: il Framework Web di Rust!
Nell’iperspazio con Rocket: il Framework Web di Rust!Nell’iperspazio con Rocket: il Framework Web di Rust!
Nell’iperspazio con Rocket: il Framework Web di Rust!Commit University
 
Vertex AI Gemini Prompt Engineering Tips
Vertex AI Gemini Prompt Engineering TipsVertex AI Gemini Prompt Engineering Tips
Vertex AI Gemini Prompt Engineering TipsMiki Katsuragi
 
APIForce Zurich 5 April Automation LPDG
APIForce Zurich 5 April  Automation LPDGAPIForce Zurich 5 April  Automation LPDG
APIForce Zurich 5 April Automation LPDGMarianaLemus7
 
Pigging Solutions in Pet Food Manufacturing
Pigging Solutions in Pet Food ManufacturingPigging Solutions in Pet Food Manufacturing
Pigging Solutions in Pet Food ManufacturingPigging Solutions
 
Automating Business Process via MuleSoft Composer | Bangalore MuleSoft Meetup...
Automating Business Process via MuleSoft Composer | Bangalore MuleSoft Meetup...Automating Business Process via MuleSoft Composer | Bangalore MuleSoft Meetup...
Automating Business Process via MuleSoft Composer | Bangalore MuleSoft Meetup...shyamraj55
 
Connect Wave/ connectwave Pitch Deck Presentation
Connect Wave/ connectwave Pitch Deck PresentationConnect Wave/ connectwave Pitch Deck Presentation
Connect Wave/ connectwave Pitch Deck PresentationSlibray Presentation
 
Unleash Your Potential - Namagunga Girls Coding Club
Unleash Your Potential - Namagunga Girls Coding ClubUnleash Your Potential - Namagunga Girls Coding Club
Unleash Your Potential - Namagunga Girls Coding ClubKalema Edgar
 
Advanced Test Driven-Development @ php[tek] 2024
Advanced Test Driven-Development @ php[tek] 2024Advanced Test Driven-Development @ php[tek] 2024
Advanced Test Driven-Development @ php[tek] 2024Scott Keck-Warren
 
Science&tech:THE INFORMATION AGE STS.pdf
Science&tech:THE INFORMATION AGE STS.pdfScience&tech:THE INFORMATION AGE STS.pdf
Science&tech:THE INFORMATION AGE STS.pdfjimielynbastida
 
SQL Database Design For Developers at php[tek] 2024
SQL Database Design For Developers at php[tek] 2024SQL Database Design For Developers at php[tek] 2024
SQL Database Design For Developers at php[tek] 2024Scott Keck-Warren
 
Bun (KitWorks Team Study 노별마루 발표 2024.4.22)
Bun (KitWorks Team Study 노별마루 발표 2024.4.22)Bun (KitWorks Team Study 노별마루 발표 2024.4.22)
Bun (KitWorks Team Study 노별마루 발표 2024.4.22)Wonjun Hwang
 
Tech-Forward - Achieving Business Readiness For Copilot in Microsoft 365
Tech-Forward - Achieving Business Readiness For Copilot in Microsoft 365Tech-Forward - Achieving Business Readiness For Copilot in Microsoft 365
Tech-Forward - Achieving Business Readiness For Copilot in Microsoft 3652toLead Limited
 
Tampa BSides - Chef's Tour of Microsoft Security Adoption Framework (SAF)
Tampa BSides - Chef's Tour of Microsoft Security Adoption Framework (SAF)Tampa BSides - Chef's Tour of Microsoft Security Adoption Framework (SAF)
Tampa BSides - Chef's Tour of Microsoft Security Adoption Framework (SAF)Mark Simos
 
Transcript: New from BookNet Canada for 2024: BNC BiblioShare - Tech Forum 2024
Transcript: New from BookNet Canada for 2024: BNC BiblioShare - Tech Forum 2024Transcript: New from BookNet Canada for 2024: BNC BiblioShare - Tech Forum 2024
Transcript: New from BookNet Canada for 2024: BNC BiblioShare - Tech Forum 2024BookNet Canada
 
WordPress Websites for Engineers: Elevate Your Brand
WordPress Websites for Engineers: Elevate Your BrandWordPress Websites for Engineers: Elevate Your Brand
WordPress Websites for Engineers: Elevate Your Brandgvaughan
 

Recently uploaded (20)

Key Features Of Token Development (1).pptx
Key  Features Of Token  Development (1).pptxKey  Features Of Token  Development (1).pptx
Key Features Of Token Development (1).pptx
 
Gen AI in Business - Global Trends Report 2024.pdf
Gen AI in Business - Global Trends Report 2024.pdfGen AI in Business - Global Trends Report 2024.pdf
Gen AI in Business - Global Trends Report 2024.pdf
 
Scanning the Internet for External Cloud Exposures via SSL Certs
Scanning the Internet for External Cloud Exposures via SSL CertsScanning the Internet for External Cloud Exposures via SSL Certs
Scanning the Internet for External Cloud Exposures via SSL Certs
 
Unraveling Multimodality with Large Language Models.pdf
Unraveling Multimodality with Large Language Models.pdfUnraveling Multimodality with Large Language Models.pdf
Unraveling Multimodality with Large Language Models.pdf
 
Nell’iperspazio con Rocket: il Framework Web di Rust!
Nell’iperspazio con Rocket: il Framework Web di Rust!Nell’iperspazio con Rocket: il Framework Web di Rust!
Nell’iperspazio con Rocket: il Framework Web di Rust!
 
Vertex AI Gemini Prompt Engineering Tips
Vertex AI Gemini Prompt Engineering TipsVertex AI Gemini Prompt Engineering Tips
Vertex AI Gemini Prompt Engineering Tips
 
APIForce Zurich 5 April Automation LPDG
APIForce Zurich 5 April  Automation LPDGAPIForce Zurich 5 April  Automation LPDG
APIForce Zurich 5 April Automation LPDG
 
Pigging Solutions in Pet Food Manufacturing
Pigging Solutions in Pet Food ManufacturingPigging Solutions in Pet Food Manufacturing
Pigging Solutions in Pet Food Manufacturing
 
Automating Business Process via MuleSoft Composer | Bangalore MuleSoft Meetup...
Automating Business Process via MuleSoft Composer | Bangalore MuleSoft Meetup...Automating Business Process via MuleSoft Composer | Bangalore MuleSoft Meetup...
Automating Business Process via MuleSoft Composer | Bangalore MuleSoft Meetup...
 
Connect Wave/ connectwave Pitch Deck Presentation
Connect Wave/ connectwave Pitch Deck PresentationConnect Wave/ connectwave Pitch Deck Presentation
Connect Wave/ connectwave Pitch Deck Presentation
 
Unleash Your Potential - Namagunga Girls Coding Club
Unleash Your Potential - Namagunga Girls Coding ClubUnleash Your Potential - Namagunga Girls Coding Club
Unleash Your Potential - Namagunga Girls Coding Club
 
Advanced Test Driven-Development @ php[tek] 2024
Advanced Test Driven-Development @ php[tek] 2024Advanced Test Driven-Development @ php[tek] 2024
Advanced Test Driven-Development @ php[tek] 2024
 
Science&tech:THE INFORMATION AGE STS.pdf
Science&tech:THE INFORMATION AGE STS.pdfScience&tech:THE INFORMATION AGE STS.pdf
Science&tech:THE INFORMATION AGE STS.pdf
 
Hot Sexy call girls in Panjabi Bagh 🔝 9953056974 🔝 Delhi escort Service
Hot Sexy call girls in Panjabi Bagh 🔝 9953056974 🔝 Delhi escort ServiceHot Sexy call girls in Panjabi Bagh 🔝 9953056974 🔝 Delhi escort Service
Hot Sexy call girls in Panjabi Bagh 🔝 9953056974 🔝 Delhi escort Service
 
SQL Database Design For Developers at php[tek] 2024
SQL Database Design For Developers at php[tek] 2024SQL Database Design For Developers at php[tek] 2024
SQL Database Design For Developers at php[tek] 2024
 
Bun (KitWorks Team Study 노별마루 발표 2024.4.22)
Bun (KitWorks Team Study 노별마루 발표 2024.4.22)Bun (KitWorks Team Study 노별마루 발표 2024.4.22)
Bun (KitWorks Team Study 노별마루 발표 2024.4.22)
 
Tech-Forward - Achieving Business Readiness For Copilot in Microsoft 365
Tech-Forward - Achieving Business Readiness For Copilot in Microsoft 365Tech-Forward - Achieving Business Readiness For Copilot in Microsoft 365
Tech-Forward - Achieving Business Readiness For Copilot in Microsoft 365
 
Tampa BSides - Chef's Tour of Microsoft Security Adoption Framework (SAF)
Tampa BSides - Chef's Tour of Microsoft Security Adoption Framework (SAF)Tampa BSides - Chef's Tour of Microsoft Security Adoption Framework (SAF)
Tampa BSides - Chef's Tour of Microsoft Security Adoption Framework (SAF)
 
Transcript: New from BookNet Canada for 2024: BNC BiblioShare - Tech Forum 2024
Transcript: New from BookNet Canada for 2024: BNC BiblioShare - Tech Forum 2024Transcript: New from BookNet Canada for 2024: BNC BiblioShare - Tech Forum 2024
Transcript: New from BookNet Canada for 2024: BNC BiblioShare - Tech Forum 2024
 
WordPress Websites for Engineers: Elevate Your Brand
WordPress Websites for Engineers: Elevate Your BrandWordPress Websites for Engineers: Elevate Your Brand
WordPress Websites for Engineers: Elevate Your Brand
 

Ethics of AI - An Engineering Perspective

  • 1. Ethics of AI An engineering perspective. Francisco Curado Teixeira (francisco.curado.teixeira@gmail.com) ECA/UN Youth Innovation Bootcamp on Emerging Technologies 23 Feb 2021 Acknowledgments: Project RETIOT, DETI-IEETA, Univ. of Aveiro (http://wiki.ieeta.pt/wiki/index.php/Institute_of_Electronics_and_Informatics_Engineering_of_Aveiro)
  • 2. Given that AI is a powerful technology, we have a moral obligation to use it well, to promote the positive aspects and avoid or mitigate the negative ones. Russell, S. J. & Norvig, P., 2020. Artificial Intelligence - A Modern Approach. Pearson ----------------- If the feedback is built into a machine that cannot be inspected until the final goal is attained, the possibilities for catastrophe are greatly increased. I should very much hate to ride on the first trial of an automobile regulated by photoelectric feedback devices, unless there were somewhere a handle by which I could take over control if I found myself driving smack into a tree. Norbert Wiener, 1964. God & Golem, Inc.
  • 3. AI Ethics by Design - a Case Study Problem Motivation.
  • 4. AI Ethics by Design - a Case Study Detection & Tracking of moving subjects in indoor / outdoor spaces. Applications (Public Surveillance & Security not contemplated) Hospitals and Healthcare ● Monitor space occupation and movement of people inside the building ● Detect and Alert for falls and inanimate people City traffic management ● Traffic assessment & route planning ● Safety of pedestrians & bicyclists: warn for imminent run-overs and collisions See Credits list: [A] See Credits list: [B] See Credits list: [D]
  • 5. AI Ethics by Design - a Case Study Ethical Considerations.
  • 6. AI Ethics by Design - a Case Study Detection & Tracking of moving subjects in indoor / outdoor spaces. Ethical and operational requirements Inclusiveness ● Account for diversity of subjects and situations ○ Humans, animals, vehicles, robots... ○ Apply to normal daily activities and detect abnormal behaviour ○ Account for people with disabilities Privacy compliance ● Avoid acquiring data that permits personal identification Safety ● Technology must be safe for humans and animals
  • 7. AI Ethics by Design - a Case Study Detection & Tracking of moving subjects in indoor / outdoor spaces. Conventional solution: cameras Advantages ● Well-established technology ● Low-cost H/W & CV software tools Problems ● Privacy violation ● Illegal in most countries See Credits list: [C]
  • 8. AI Ethics by Design - a Case Study Technical Solutions.
  • 9. Detection & Tracking of moving subjects in indoor / outdoor spaces. Nonintrusive solutions: RADAR Advantages ● Nonintrusive: no ID information in radar data ● Automotive Radar: compact, affordable, versatile FMCW tech. ● Safe for humans: 77-81 GHz ● Robust against environmental disturbances ● Appearing in mobile phones Limitations ● No direct ‘picture’ acquisition ● Feature extraction highly dependent on radar hardware AI Ethics by Design - a Case Study
  • 10. Detection & Tracking of moving subjects in indoor / outdoor spaces. Nonintrusive solutions: RADAR Radar working Principles and Techniques ● Cloud point of reflecting objects ● Rich data: range, angle, velocity, intensity... Problems ● Noisy raw data: clutter, multi-path... ● Complex feature extraction AI Ethics by Design - Case Study
  • 11. Detection & Tracking of moving subjects in indoor / outdoor spaces. Nonintrusive solutions: RADAR Radar working Principles and Techniques ● Cloud point of reflecting objects ● Rich data: range, angle, velocity, intensity... Problems ● Noisy raw data: clutter, multi-path... ● Complex feature extraction AI Ethics by Design - Case Study
  • 12. Detection & Tracking of moving subjects in indoor / outdoor spaces. Nonintrusive solutions: RADAR Radar working Principles and Techniques ● Cloud point of reflecting objects ● Rich data: range, velocity, intensity... Problems ● Noisy raw data: clutter, multi-path... ● Complex feature extraction AI Ethics by Design - Case Study
  • 13. Detection & Tracking of moving subjects in indoor / outdoor spaces. Nonintrusive solutions: RADAR Motion detection Principles and Techniques ● Point cloud data: range, angle, Doppler, Intensity… ● Cluster points: range & Doppler (velocity) ● Use Doppler patterns for classification ● Complementary: signal intensity (RCS) AI Ethics by Design - Case Study Radar geometry in 2D. Source: [4] Illustration of a person walking and a robot moving and the respective point clouds. Source: [5]
  • 14. Detection & Tracking of moving subjects in indoor / outdoor spaces. Nonintrusive solutions: RADAR Complexity ● Advanced signal processing (raw data): ADC, 3D FFT, Clustering, Data- Association, Tracking… ● Novel feature extraction techniques Advantages ● Diversity of available feature types (e.g.: RCS) and dimensionality (e.g. multi-frame) AI Ethics by Design - Case Study Doppler pattern of mannequin or robot. Source: [5] Doppler pattern of person walking. Source: [5]
  • 15. AI Ethics by Design - a Case Study Detection & Tracking of moving subjects in indoor / outdoor spaces. Alternative solution: Low-resolution Thermal camera Advantages ● Nonintrusive: no personal ID information ● Well-established, affordable technology ● Compact Limitations ● Sensitivity to environmental disturbances ● Controversial in public spaces: similarity to Video & 3D cameras
  • 16. AI Ethics by Design - a Case Study Adopted Solution.
  • 17. AI Ethics by Design - a Case Study Anonymized Detection & Tracking of moving subjects in indoor / outdoor spaces. Preferred solution: RADAR-only sensing Advantages of Radar Data ● Subject classification based on velocity (Doppler) patterns ● Multi-class classification: pedestrians, bicyclists, cars, animals... Problems: ● Limited availability of AI tools and scarcity of labeled data-sets ● Huge amount of data supplied by the radar ● The majority of radar frames consist of noise and requires human inspection ● Radar frames are NOT easily interpretation by humans ● Data labeling is a highly time-consuming, expensive and tedious process ?? *!* ?!
  • 18. AI Ethics by Design - a Case Study Anonymized Detection & Tracking of moving subjects in indoor / outdoor spaces. Proposed: Training with Radar + Thermal data Data acquisition and Labeling: Initial setup ● Radar board + Raspberry Pi CPU + camera ● WiFi + PC ● Using camera images to select radar frames allows for automatic or supervised labeling of subjects Final setup ● Low-res, low-cost thermal camera replacing RPi camera
  • 19. AI Ethics by Design - a Case Study Anonymized Detection & Tracking of moving subjects in indoor / outdoor spaces. Proposed: Training with Radar + Thermal data Deployment in Laboratorial / Academic environment ● Nonintrusive: no personal ID information ● Inform the community about the Experiment Advantages ● Favorable context for system deployment ● Radar and Camera can be left shooting without supervision
  • 20. AI Ethics by Design - a Case Study Results.
  • 21. Anonymized Detection & Tracking of moving subjects in indoor / outdoor spaces. Model Learning: RADAR-only deployed on a mobile robot Multi-class Classification ● People ● Mannequin ● Person in wheelchair ● Robots AI Ethics by Design - a Case Study
  • 22. Anonymized Detection & Tracking of moving subjects in indoor / outdoor spaces. Model Learning: RADAR-only deployed on a mobile robot Results with initial datasets (25431 samples) after hyper-parameter optimization ● Accuracy: Artificial Neural Network (ANN) ○ Doppler only: 92% ○ Doppler + RCS: 96% ● Accuracy: Random Forest (RF) ○ Doppler only: 92% ○ Doppler + RCS: 95% AI Ethics by Design - a Case Study
  • 23. Anonymized Detection & Tracking of moving subjects in indoor / outdoor spaces. Model Learning: RADAR-only sensing Increasing the dataset ● Start with a small initial labeled dataset and trained model (classifier) ● Apply Active and Cooperative Learning to enlarge the labeled dataset and improve the model ● Use synthetic or augmented data: physical models; stochastic data-augmentation AI Ethics by Design - a Case Study Active Learning loop
  • 24. AI Ethics by Design - a Case Study Conclusions.
  • 25. Anonymized Detection & Tracking of moving subjects in indoor / outdoor spaces. Anonymized detection, classification, and tracking ● It is possible to implement this type of system without using vision ● It is crucial to address ethical issues in the design phase ● Ethical considerations may condition the choice of H/W and S/W Tools ● Alternative technologies and methods may present increased complexity and cost but may reveal more robust and versatile AI Ethics by Design - a Case Study
  • 26. Thank you Acknowledgments: Project RETIOT, DETI-IEETA, Univ. of Aveiro (http://wiki.ieeta.pt/wiki/index.php/Institute_of_Electronics_and_Informatics_Engineering_of_Aveiro)
  • 27. References 1. Russell, S. J. & Norvig, P., 2020. Artificial Intelligence. - A Modern Approach. Pearson 2. Dubber, M. D.; Pasquale, F. & Das, S. (Eds.), 2020. The Oxford Handbook of ETHICS OF AI. Oxford University Press 3. Ng, Andrew, 2018. Machine Learning Yearning. www.dbooks.org 4. Livshitz, M. 2018. Tracking radar targets with multiple reflection points. https://e2e.ti.com/cfs- file/__key/communityserver-discussions-components-files/1023/ [Online accessed 13-June-2019]. 5. Castanheira, J., 2019. Machine Learning Methods for Radar-Based People Detection. Master thesis. Universidade de Aveiro. Figures’ credits A. By Micov at English Wikipedia, CC BY-SA 3.0: https://commons.wikimedia.org/w/index.php?curid=47564675 B. Dllu: https://commons.wikimedia.org/wiki/User:Dllu C. Paweł Zdziarski, CC BY-SA 3.0 via Wikimedia Commons <http://creativecommons.org/licenses/by-sa/3.0/> D. US News - Health: https://health.usnews.com/health-news/patient-advice/slideshows/14-ways-to-protect-seniors-from-falls