Professional Documents
Culture Documents
Module 4-6-Training For AI Ethics - Bangkok & Indonesia Feb Mar 2023 - Final
Module 4-6-Training For AI Ethics - Bangkok & Indonesia Feb Mar 2023 - Final
Module 4-6-Training For AI Ethics - Bangkok & Indonesia Feb Mar 2023 - Final
© Copyrighted to IEEE
AGENDA
• How Standards support Critical and Emerging Technology, especially Artificial Intelligence (AI)
• Global Standardization on AI
• Gender Responsive Standard & How to develop
© Copyrighted to IEEE
IEEE – OVER 135 YEARS OF INNOVATION
4IR
3IR
2IR
3
© Copyrighted to IEEE
ADVANCING
TECHNOLOGY
FOR HUMANITY
ABOUT IEEE
▪ Inspiring a global community of innovation
▪ Where forward-thinking professionals collaborate to:
▪ Discover what’s next in tech innovation
▪ Build technical communities
▪ Shape and share research
▪ Engage in Humanitarian activities
▪ Create global standards
4
© Copyrighted to IEEE
RAISING THE WORLD’S STANDARDS WHAT STANDARDS DO FOR INDUSTRY & SOCIETY
Ensure system Continue to
interoperability enhance technology
while enabling product by motivating participation,
differentiation collaboration & competition
from a diverse set of companies
© Copyrighted to IEEE
RAISING
THE WORLD’S
STANDARDS
ABOUT IEEE SA
Developing market relevant open standards and solutions:
▪ Advancing global technologies and technology platforms
▪ Promoting innovation
▪ Protecting public safety, health & wellbeing
▪ Contributing to a sustainable future
© Copyrighted to IEEE
AI LANDSCAPE
© Copyrighted to IEEE
Digital transformation seems to be at the forefront
of every business’s organizational strategy.
It is defined as “the adoption of digital technology
to transform services or businesses, through
replacing non-digital or manual processes with
digital processes or replacing older digital
technology with newer digital technology”.
-ServicePath Website based on research from Accenture, Deloitte, and Mckinsey
https://servicepath.co/2021/04/the-future-of-digital-transformation-accenture-deloitte-and-mckinsey/#:~:text=Digital%20transformation%20should%20be%20at,digital%20technology%20with%20newer%20digital
© Copyrighted to IEEE
THE USE OF AI SYSTEMS IS ONLY GROWING ….
• Worldwide revenues for the artificial intelligence (AI) market,
including software, hardware, and services, are forecast to grow
year over year in 2021 to ~ $1,800B by 2030
• There is a conscious increase by services providers to applying AI to
solve industry- and domain-specific problems for clients
• ~80M US Adults use an AI assistant in their cars at least once a
Market Growth
month
• 50+% of companies monitor AI-created bias
• Regional and National strategies have only increased investments
into horizontal and vertical-oriented AI systems – coordination
across University systems and research bodies is key.
© Copyrighted to IEEE
© Copyrighted to IEEE
© Copyrighted to IEEE
© Copyrighted to IEEE
© Copyrighted to IEEE
© Copyrighted to IEEE
© Copyrighted to IEEE
© Copyrighted to IEEE
© Copyrighted to IEEE
© Copyrighted to IEEE
© Copyrighted to IEEE
“…meeting the needs of the present without
compromising the ability of future generations to
meet their own needs.”
© Copyrighted to IEEE
REAL, UNPREDICTABLE CHALLENGES
Copyright Issues
Algorithmic Equity Imbalance Algorithmic Bias Algorithmic Cheating
© Copyrighted to IEEE
AND THESE ARE NOT ONE OFF EXAMPLES…
© Copyrighted to IEEE
DIGITAL TRANSFORMATION JOURNEY - ELEMENTS FOR SUCCESS
© Copyrighted to IEEE
NOTED AI RISKS
• Given the broad reach and impact of AI
technologies and associated markets,
the need for regulatory compliance
presents a natural risk concern.
• Socio-technical elements make up 33%
of the categories associated with AI
Market Perspectives
© Copyrighted to IEEE
… INCREASING MARKET DRIVERS THAT DEMAND ONGOING,
DEMONSTRATED TRUSTWORTHY ENGAGEMENTS AT MULTIPLE LEVELS &
TOUCHPOINTS…
• Soft or socio-technical attributes are difficult to demonstrate compliance to the marketplace
without accepted and proven certification schemes from recognized entities
• AI Ethics expertise available to large companies is not necessarily available to small and mid-
size enterprises; increasing the competitive gaps and challenging resource prioritization to
earn compliance where necessary
Market Considerations
• Research has shown bias in AI-based systems producing detrimental results to people of color,
gender, and more. The negative brand perception and impact can lead to loss of revenues and
business. Further, the trusted relationship with an Entity’s stakeholders are not easily repaired.
• Draft regional regulations seek to encourage socio-technical considerations to be
implemented in situations where the risk is unacceptable or high-risk. Yet, there is no evident
test or standardized approach that Entity’s can rely upon immediately without undertaking an
additional risk of further testing fees or AI solutions being deployed with a risk of having to
recall.
• Cities and Governments tasked with procuring technologies with embedded or hosted AI
systems require the means to properly vet the ethical AI concerns and considerations to
ensure citizens and residents are treated consistently in the most trust-worthy manner
possible.
© Copyrighted to IEEE
“Our industry has started to embrace and apply key “We want to design healthy relationships with our users. The
technologies such as AI and robotics, which could potentially potential of AI is wrapped up in its longevity as a solution—
disrupt the whole energy value chain. We should be thoughtful meaning everything we design must address current and
about delivering this change in the right way, and avoiding future needs for users. To truly understand those needs, we
unintended consequences, before the technology becomes need an inclusive and ethical approach to the entire
widespread.” process. Globally, we are starting to see the repercussions that
come when companies do not prioritize AI ethics in their
solutions. We want to make sure that ethical practices are
“Our core values drive everything we do. Trust,
ingrained on our teams so they can then be embedded into
customer success, innovation, and equality
the products themselves.”
mean that developing AI our customers and
society can trust is critical to our mission.”
“…. a values-driven company who puts our responsibilities toward kids and parents at the
center of our decision making. We aim to apply the same rigor to our implementation of
technical experiences/ solutions as we do to production, quality, and safety.”
“We see opportunities for AI and automation to be embedded into our day-today technology design and operational processes, so it’s
vital that our people are aware of and understand AI ethics in the context of their work. This will enable us to proactively
manage potential impacts on employees, customers, and society.”
“We have an ethical foundation which we continuously work to strengthen and expand. In the past this has included adopting a
progressive code of conduct, developing and adhering to global human rights principles, and assessing our supply chain to enable
sustainable development including pursuing conflict-free minerals. As we become more data and AI-centric, we see the need to
expand our principles in such a way to mitigate any future harm our hardware, software, or data science contributions could
make in the world.”
© Copyrighted to IEEE
Source: EAD for Business
OBSERVED TECHNIQUES & APPROACHES TOWARDS TRUSTWORTHY AI
Increased Hiring of Roles Dedicated to AI Ethics
• Chief AI Ethics Officer
• Data Ethicists
• Responsible AI Ethicist
© Copyrighted to IEEE
Principles & Frameworks
Tools
Standards
Certification
AI Literacy
Market Basis
Community
AI Governance Platforms
© Copyrighted to IEEE
EAD for Business:
AI Ethics Readiness Framework
© Copyrighted to IEEE
TECHNOLOGY ETHICS
© Copyrighted to IEEE
“Salesforce’s core values drive everything we do. Trust, customer success, innovation, and
equality mean that developing AI our customers and society can trust is critical to our
mission.”
“We have an ethical foundation which we continuously work to strengthen and expand. In the past this has
included adopting a progressive code of conduct, developing and adhering to global human rights
principles, and assessing our supply chain to enable sustainable development including pursuing conflict-
free minerals. As we become more data and AI-centric, we see the need to expand our principles in such a
way to mitigate any future harm our hardware, software, or data science contributions could make in the
world.” — Anna Bethke, Intel
“We see opportunities for AI and automation to be embedded into our day-today technology design and
operational processes, so it’s vital that our people are aware of and understand AI ethics in the context of
their work. This will enable us to proactively manage potential impacts on employees, customers, and
society.”
Source: EAD for Business —Amy Oding, Vodafone
© Copyrighted to IEEE
“Our industry has started to embrace and apply key technologies such as AI and robotics,
which could potentially disrupt the whole energy value chain. We should be thoughtful about
delivering this change in the right way, and avoiding unintended consequences, before
—Bob Flint, BP
the technology becomes widespread.”
“We want to design healthy relationships with our users. The potential of AI is wrapped up in
its longevity as a solution—meaning everything we design must address current and future
needs for users. To truly understand those needs, we need an inclusive and ethical approach
to the entire process. Globally, we are starting to see the repercussions that come when
companies do not prioritize AI ethics in their solutions. We want to make sure that ethical
practices are ingrained on our teams so they can then be embedded into the products
themselves.”
—Milena Pribić, IBM
“The LEGO Group is a values-driven company who puts our responsibilities toward kids and
parents at the center of our decision making. We aim to apply the same rigor to our
implementation of technical experiences/ solutions as we do to production, quality, and
safety.”
© Copyrighted to IEEE
AS INVESTMENTS IN AI INNOVATION & USE INCREASE, RISK MITIGATION HAS NOT MATCHED…
Source: https://www.mckinsey.com/capabilities/quantumblack/our-insights/the-state-of-ai-in-2022-and-a-half-decade-in-review
© Copyrighted to IEEE
A Deloitte global survey of 500 government leaders in 2021
found that 92 percent of respondents at the federal level,
95 percent at the state level, and 84 percent at the local
level are of the view that AI is “mission critical” over the
next five years.[4] At least half of those surveyed listed gaps
in skills as a crucial reason for the inability of governments
to utilize AI applications in the most effective manner.
Source: https://www.orfonline.org/research/g20-ai-national-strategies-global-ambitions/
© Copyrighted to IEEE 35
FURTHER EMPHASIZING THE NEED FOR POLICIES EMERGING ENCOURAGING GREATER
TRUSTWORTHY IMPLEMENTATION
© Copyrighted to IEEE
AI POLICY PERSPECTIVES: DIFFERENT APPROACHES REFLECTED IN NATIONAL AI STRATEGIES
Citation: TY - JOUR, AU - Paunov, Caroline ,AU - Planes-Satorra, Sandra, AU - Ravelli, Greta, PY - 2019/10/17, SP - , T1 - REVIEW OF https://www.researchgate.net/figure/Timeline-of-national-adoption-of-AI-strategies_fig1_336686040
NATIONAL POLICY INITIATIVES IN SUPPORT OF DIGITAL AND AI-DRIVEN INNOVATION, DO - 10.1787/15491174-en, ER - Source: https://www.plattform-lernende-systeme.de/ai-strategies.html
© Copyrighted to IEEE
GLOBAL AI STANDARDIZATION
© Copyrighted to IEEE
EXAMPLE: POLICY & REGULATORY PURSUIT OF
MANAGING TRUSTWORTHY AI
IMPLEMENTATIONS WITH THE DRAFT EU AI ACT
© Copyrighted to IEEE
PROPOSAL FOR EU HARMONIZED RULES ON AI
© Copyrighted to IEEE
Risk
Management
EU AI Act: High Risk System Requirements
Accuracy,
Data
Robustness &
Governance
Cybersecurity
Transparency &
Provision of
Record Keeping
Information to
Users
© Copyrighted to IEEE
EU AI Act - High Risk AI Systems
Broad Definition of AI Risk Based Approach
Systems (ML + Logic (Fundamental Human
Based+ Statistical Rights, Human Mental
Models) and Physical Safety)
Policies Impacting AI System
European AI Act Elements
Compliance
Categorization
Market Intended AI
Requirements
System Use
(Unacceptable, High,
Not-High, Minimal)
Internal Assessments,
Value Chain
Conformity Assessment
Responsibilities and
(ex-ante) & Monitoring
Obligations
(Post-market)
© Copyrighted to IEEE
EU AI Act - High Risk AI Systems
Transparency &
Technical Accuracy, Robustness
Risk Management Data Governance Record Keeping Provision of Human Oversight
Documentation & Cybersecurity
Information to Users
Policies Impacting AI System
European AI Act Elements
© Copyrighted to IEEE
DIFFERENT WAYS TO APPROACH (RISK VS. IMPACT)
Canadian Impact
EU AI Act AI Bill of Rights
Assessment Act
© Copyrighted to IEEE
STANDARDIZATION
© Copyrighted to IEEE
INFLUENCED THE HUMAN FOCUSED AI / TECHNICAL S LANDSCAPE
AI ETHICS P7000 SERIES ASILOMAR AI ETHICS GUIDELINES IEEE 2863 IEEE 7000-2021
INITIATIVE OCEANIS Recommended IEEE Standard Model
LAUNCHES PRINCIPLES EU High Level
IEEE IEEE Practice for Process for
IEEE Standards Working FLI Institute Experts Group
Work Begins July 2018 Organizational Addressing Ethical
Groups January 2017 April 2019 Governance of
October 2015 Inspired by EAD Concerns During
Artificial Intelligence
June 2016 Inspired by EAD System Design
AI PRINCIPLES
February 2020 LAUNCHED 2021
OECD
May 2019
IEEE 7010-2020
Recommended
Practice for Assessing
EVERYDAY ETHICS FOR AI the Impact of
IBM Autonomous and
© Copyrighted to IEEE October 2019 Intelligent Systems on
Human Well-Being
LAUNCHED 2020
• European Commission
Collaborations and Engagements
• Council on Europe
• OECD
• UNESCO
• National initiatives
• Cities
• Standardization organizations
• Industry associations
• NGOs, Consumers..
• Global Initiative
• Council for Extended Intelligence (CXI)
• Open Community for Ethics in
Autonomous and Intelligent Systems (OCEANIS)
• Digital Inclusion and Trust in Agency
• …
© Copyrighted to IEEE
ETHICALLY ALIGNED DESIGN
Embargoed Version
AIS ETHICS IN ACTION SINCE 2015
ETHICALLY
ALIGNEDDESIGN
First Edition
A Vision for Prioritizing Human Well-being
with Autonomous and Intelligent Systems
• First version released in 2016. • EAD,v2 used by OECD for their AI principles.
• Latest version released March, 2019. • EAD, v2 used by IBM for “Everyday Ethics for AI.”
• Received over 500 pages of feedback. • EAD, v2 used by FLI for their AI Principles.
• Written by more than 700 global thought leaders. • EAD, v2 used by UNICEF for their Children’s Data principles.
• Over 300 pages long – not just a list of AI Principles. • EAD, v2 used by UNESCO for their AI Principles.
• Features hundreds of evergreen, pragmatic recommendations.
• EAD in all versions mentioned in more than three dozen academic journals, AI
• Inspired IEEE’s AIS Ethics Certification work. Principles and media since 2016
© Copyrighted to IEEE
IEEE SA IMPACT STANDARDS
• IEEE 7000-2021™ - Standard for Model Process for Addressing • IEEE 7010-2020™ - IEEE Recommended Practice for Assessing the
Ethical Concerns During System Design Impact of Autonomous and Intelligent Systems on Human Well-being
• IEEE 7001-2021™ - Standards for Transparency of Autonomous • IEEE P7010.1™ - IEEE Recommended Practice for Environmental Social
Systems Governance (ESG) and Social Development Goal (SDG) Action
Implementation and Advancing Corporate Social Responsibility
• IEEE 7002-2022™ - Standard for Data Privacy Process
• IEEE P7011™ - Standard for the Process of Identifying and Rating the
• IEEE P7003™ - Standard for Algorithmic Bias Considerations Trustworthiness of News Sources
• IEEE P7004™ - Standard for Child and Student Data Governance • IEEE P7012™ - Standard for Machine Readable Personal Privacy Terms
• IEEE P7004.1™ - Recommended Practices for Virtual Classroom • IEEE P7014™ - Standard for Ethical considerations in Emulated
Security, Privacy and Data Governance Empathy in Autonomous and Intelligent Systems
STANDARDS
• IEEE 7005-2021™ - Standard for Transparent Employer Data • IEEE P7015™ - , Standard for Data and Artificial Intelligence (AI)
Governance Literacy, Skills, and Readiness
• IEEE 7007-2021™ - Ontological Standard for Ethically Driven • IEEE P7016™ - Standard for Ethically Aligned Design and Operation of
Robotics and Automation Systems Metaverse Systems
• IEEE P7008™ - Standard for Ethically Driven Nudging for Robotic, • IEEE 2089-2021™ - IEEE Standard for an Age Appropriate Digital
Intelligent and Autonomous Systems Services Framework Based on the 5Rights Principles for Children
• IEEE P7009™ - Standard for Fail-Safe Design of Autonomous and • IEEE P2863™-Recommended Practice for Organizational Governance
Semi-Autonomous Systems of Artificial Intelligence
© Copyrighted to IEEE
VALUES ORIENTED, METRICS-DRIVEN TECHNOLOGY AT THE OUTSET OF DESIGN
Ethical Concerns & Systems Design
© Copyrighted to IEEE
VALUES ELICITATION = DESIGN REQUIREMENTS
IEEE 7000-2021
© Copyrighted to IEEE
OVERVIEW
IEEE 7000TM
1 What it delivers
Who is it for Development Stage
4 Ethical Requirements
Risk-based design
© Copyrighted to IEEE
IEEE 7000TM – WHAT IT DELIVERS
© Copyrighted to IEEE
IEEE 7000TM – WHO IS IT FOR
© Copyrighted to IEEE
IEEE 7000TM – VALUE-BASED SYSTEM DESIGN
© Copyrighted to IEEE
STAGE ONE - CONCEPT EXPLORATION
© Copyrighted to IEEE
STEP ONE - CONCEPT OF OPERATIONS
• Future focused
• At scale impact
© Copyrighted to IEEE 58
STEP TWO - EXPLORING ETHICAL VALUES
What are…
• The values that may influence the requirements and the
design of the SOI Concept Exploration Stage Development Stage
• The decisions made by the extended design team that will
foster positive social values and avoid negative values Concept of
Ethical Values Ethical
Operations & Ethical Risk-
Elicitation & Requirements
Context Based Design
Prioritization Definition
Exploration Process
Value Lead Process
Process Process
© Copyrighted to IEEE 59
ETHICAL VALUES ELICITATION & PRIORITIZATION
© Copyrighted to IEEE 60
ELICITING VALUES – QUESTIONS TO ASK
Three fundamental ethical theories provide the core questions the Value lead asks for the
project team to determine the values the SOI is to have designed into it
Utilitarian ethics
• “What benefits or harms would arise when the SOI is deployed?“
• How can the SOI foster positive benefits or promote negative harms
Virtue Ethics
• “What are the effects of the system on the character of the User?”
• How could the SOI or its features nurture positive characteristics or promote vices
Duty ethics
• “What are the personal principles of the User which can be undermined or fostered by the
SOI?”
• How the day-to-day personal values that could be supported or undermined
• Duty ethics is also important for prioritizing values identified
© Copyrighted to IEEE 61
STAGE TWO - DEVELOPMENT STAGE
© Copyrighted to IEEE 62
STEP THREE - ETHICAL VALUE REQUIREMENTS
© Copyrighted to IEEE 63
STEP FOUR - ETHICAL RISK-BASED DESIGN
© Copyrighted to IEEE 64
TRANSPARENCY MANAGEMENT
© Copyrighted to IEEE 65
IEEE 7000TM AND SYSTEM DEVELOPMENT
© Copyrighted to IEEE 66
IEEE 7001TM - IEEE STANDARD FOR TRANSPARENCY OF AUTONOMOUS SYSTEMS
Purpose:
The purpose of this standard is to set out measurable, testable levels of transparency
for autonomous systems. The general principle behind this standard is that it should
always be possible to understand why and how the system behaved the way it did.
Transparency is one of the eight General Principles set out in IEEE Ethically Aligned
Design [B21], stated as “The basis of a particular autonomous and intelligent system
decision should always be discoverable.”
© Copyrighted to IEEE 67
IEEE 7002TM - IEEE STANDARD FOR DATA PRIVACY PROCESS
© Copyrighted to IEEE 68
IEEE 7005TM - IEEE STANDARD FOR TRANSPARENT EMPLOYER DATA GOVERNANCE
© Copyrighted to IEEE 69
IEEE 7007TM - IEEE ONTOLOGICAL STANDARD FOR ETHICALLY DRIVEN ROBOTICS AND
AUTOMATION SYSTEMS
A set of ontologies with different abstraction levels that contain concepts, definitions,
axioms, and use cases that assist in the development of ethically driven methodologies
for the design of robots and automation systems is established by this standard. It
focuses on the robotics and automation domain without considering any particular
applications and can be used in multiple ways, for instance, during the development of
robotics and automation systems as a guideline or as a reference “taxonomy” to enable
clear and precise communication among members from different communities that
include robotics and automation, ethics, and correlated areas. Users of this standard
need to have a minimal knowledge of formal logics to understand the axiomatization
expressed in Common Logic Interchange Format.
© Copyrighted to IEEE 70
IEEE 7010-2020: Societal Impact Assessment Framework
7010-2020 - IEEE
Recommended Practice for
Assessing the Impact of
Autonomous and Intelligent
Systems on Human Well-Being
• 7010-2020 helps AI Systems creators and policymakers to ask the critical question:
What are we optimizing our AI Systems to do as Key Performance Indicators of
societal success?
IEEE 7010-2020
https://ieeexplore.ieee.org/document/9084219
© Copyrighted to IEEE
IEEE 2089TM - IEEE STANDARD FOR AN AGE APPROPRIATE DIGITAL SERVICES
FRAMEWORK BASED ON THE 5RIGHTS PRINCIPLES FOR CHILDREN
This standard is the first in a family of standards focused on the 5Rights principles and
establishes a set of processes for developing age appropriate digital services for
situations where users are children. The framework centers around the following key
areas as follows:
a) Recognition that the user is a child
b) Consideration for the capacity of and upholds the rights of children
c) Offers terms appropriate to children
d) Presents information in an age appropriate way
e) Offers a level of validation for service design decisions
© Copyrighted to IEEE 72
IEEE 2089TM - IEEE STANDARD FOR AN AGE APPROPRIATE DIGITAL SERVICES
FRAMEWORK BASED ON THE 5RIGHTS PRINCIPLES FOR CHILDREN
© Copyrighted to IEEE 73
IEEE P2863 Recommended Practice for Organizational
Governance of Artificial Intelligence aims to specify governance
criteria such as safety, transparency, accountability,
AI SYSTEMS GOVERNANCE /
© Copyrighted to IEEE
PRINCIPLES TO PRACTICE+
A SYSTEMATIC APPROACH
Principles
RAISING THE WORLD’S
Standards Market Acknowledgement Ecosystem
RAISING THE WORLD’S STANDARDS
STANDARDS
• Training
• Assessor
• Audit
• Registry
• Strategic
Partners
• DataPort
76 © Copyrighted to IEEE
11
DEEP DIVE: AIE SYSTEM CERTIFICATION
© Copyrighted to IEEE
Step 1- Information about the system (ConOps Demo)
Five Steps to IEEE AI Ethics Assessment of ECS
© Copyrighted to IEEE
Step 2- Ethics Profiling & Impact Assessment
Five Steps to IEEE AI Ethics Assessment of ECS
© Copyrighted to IEEE
Five Steps to IEEE AI Ethics Assessment of ECS Step 3- The Selection of Pertinent Ethics Suite & Criteria
Based on the information an ethics profiling and
impact assessment according to the IEEE AI Ethics
Certification framework was conducted.
These criteria range from rather technical
aspects such as;
■ Error analysis
■ Hyperparameter tuning
■ Mitigation of false positives
© Copyrighted to IEEE
Five Steps to IEEE AI Ethics Assessment of ECS Step 4 - The Case for Ethics document
For each of the criteria, client provided evidence in
the form of
■ Technical documentation
■ system architecture and software implementation details
■ process and role definitions
■ organigrams, etc.
© Copyrighted to IEEE
Five Steps to IEEE AI Ethics Assessment of ECS Step 5 – Assessment Report
Finally, an Assessment & Ethics Progressions Report
was delivered back to the client by IEEE SA. This
included
■ Specific feedback for each of the criteria from
the expert panel members, indicating to what
degree the respective criterion was considered
fulfilled
■ What could be done to further improve in the
respective areas
■ An overall confirmation that the submitted Case for
Ethics justifies recognition and certification through
the IEEE AI Ethics Certification program for client’s
ECS
Source: https://beyondstandards.ieee.org/the-ieee-certifaied-framework-for-ai-ethics-applied-to-the-city-of-vienna/
© Copyrighted to IEEE
IEEE CertifAIEd - Five Steps to IEEE AI Ethics Assessment of ECS
© Copyrighted to IEEE
EXAMPLES IN ACTION
© Copyrighted to IEEE
In action – City of Vienna
© Copyrighted to IEEE
IN THE NEWS
Problem set/ rationale
AI Ethics requires all parts of the organization to have a base level of AI
literacy and alignment. From procurement to
Second, senior leaders often don’t have the requisite knowledge for
spotting ethical flaws in their organization’s AI, putting the company at
risk, both reputationally and legally
Third, an AI ethical risk program requires knowledgeable data scientists
and engineers
What they did about it
Real World Example
© Copyrighted to IEEE
For reflection:
© Copyrighted to IEEE
Governments are not immune from Algorithmic Bias
© Copyrighted to IEEE
Amplification of Miscarriages of Justice
© Copyrighted to IEEE
© Copyrighted to IEEE
● Key items from the paper
1. Establish a whole society view and mapping of the broader goals we seek to
achieve;
© Copyrighted to IEEE
© Copyrighted to IEEE
● Key reflections and cross-learnings for AI from the paper
1. Highlights the importance of a holistic understanding to policy making.
2. It is not sufficient to have “gender blind” policies - Policies are incorrectly
considered neutral by policy makers where they ignore the differential impacts
they have on different genders and socioeconomic and cultural groups. Based on
such false assumptions, the policies are less effective and/or have unintended
effects. Gender awareness and sensitive not homogenous and equal
approaches. Herein lies the difference between equality and equity.
3. Recognises the important role that energy, energy infrastructure, connectivity
and digital infrastructure as well as access to the internet play with regard to any
government being able to fulfil on its National AI strategies.
4. The availability and use of energy and domestic appliances in the home tends to
impact women more than men, and can be a barrier to paid for work, education
and training outside the home.
5. Energy, energy infrastructure, digital infrastructure, connectivity and digital
access through devices and broadband availability presents a dependency
potentially on private sector to act where these are not nationalised industry.
© Copyrighted to IEEE
According to a report commissioned by the European Parliament’s
Committee on Women’s Rights and Gender Equality:
Ref: Women, Gender Equality and the Energy Transition in the EU (europa.eu)
Dated May 2019
The DIITA report shows that energy policies which neglect gender issues and
also fail to take into account other social characteristics, such as age, marital
status, and ethnicity, which also influence energy access, and can
disproportionately affect one gender over another.
© Copyrighted to IEEE
© Copyrighted to IEEE
● Key reflections from the paper
● Gender stereotypes can be embedded in avatars
● AR/XR presents new challenges to privacy
● Information shared by users can be used by providers to infer information about
gender.
● The availability of the sensed user data, coupled with real-time local and edge
computing processing and sensor fusion, introduces and amplifies not only risks
to anonymity and privacy, but also identity.
● Behaviour, including hand - eye tracked activity and, internal mental states are
predicted or inferred or linked by AI.
● Inferences and insight from AI are being used to classify and profile gender or
sexual orientation or sexual preferences!
● Such inferences based on unverified or inaccurate data can falsely identify a
woman as a man (false positive) or incorrectly not identify a person as a woman
at all (false negative).
© Copyrighted to IEEE
● Key Takeaways for Gender Based AI Ethics Policy Making
● Awareness - increase awareness of the potential for AI to discriminate and cause
unfair outcomes, and understanding about how standards and an outcomes
based approach can help identify these AI ethics risks.
● Intelligence - Upskill your government departments, legislators, regulators and
businesses to AI ethics issues, to apply an outcomes-based approach to policy
making and by engaging new and emerging AI applications in ways which include
multiple stakeholders such as in a Regulatory Sandbox
● Expertise
a. Consider becoming the first South East Asian nation to be IEEE CertifAIEd
b. Mandate the use of AI Ethics standards in government (procurement)
contracts
c. Promote AI Ethics standards use amongst businesses to increase
international competitiveness
d. Mandate compliance with certain ethics guidelines, international AI ethics
guidelines © Copyrighted to IEEE
● Remember: IEEE are here to help you on your AI Ethics journey
ADDITIONAL RESOURCES TO SUPPORT YOU ON
YOUR JOURNEY
© Copyrighted to IEEE
IEEE GET AI Ethics & Governance Standards Program
Purpose: List of Standards*:
IEEE SA, together with some strategic partners makes AIE & 2089-20212089-2021 - IEEE Standard for an Age
Governance standards available at no cost to help raise awareness Appropriate Digital Services Framework Based on the
5Rights Principles for Children
and literacy around AI ethics and help AI developers incorporate
7000-20217000-2021 - IEEE Standard Model Process for
human-centric design principles into their product roadmaps and Addressing Ethical Concerns during System Design
organizational and governance structures, supporting more trustworthy 7001-20217001-2021 - IEEE Standard for Transparency of
AI systems. Autonomous Systems
7002-20227002-2022 - IEEE Standard for Data Privacy
Terms of Use: Process
7005-20217005-2021 - IEEE Standard for Transparent
-Access, download, print and/or retain one (1) copy of each Document
Employer Data Governance
in this Program 7007-20217007-2021 - IEEE Ontological Standard for
for individual use only, including job-related functions. Ethically Driven Robotics and Automation Systems
-Not to further copy, prepare, and/or distribute copies of the Document, 7010-20207010-2020 - IEEE Recommended Practice for
nor significant portions of the Document, in any form, without prior Assessing the Impact of Autonomous and Intelligent
written permission from IEEE. Systems on Human Well-Being
-Contact us at AIE-GET@ieee.org for further information
*More to be added as they are approved.
© Copyrighted to IEEE
AIE READINESS FRAMEWORK
© Copyrighted to IEEE
AI LITERACY THROUGH THE IEEE AIE FRAMEWORK
Awareness Intelligence Expertise
A
•
•
•
Introduction to AI Ethics
AI Standards: Roadmap for Ethical and
Responsible Digital Environments
Artificial Intelligence (AI) Assurance Artificial
Intelligence and Ethics in Design: Responsible
•
•
•
•
•
•
I
Certification
Transparency
Accountability
Algorithmic Bias
Privacy
Enterprise Governance
•
•
•
•
E+
AI Ethics CertifAIEd Assessor course
CertifAIEd examination
CertifAIEd formal review
CertifAIEd designation
• Custom Workshops &
Training to meet the
needs of your
organization
© Copyrighted to IEEE
TRUSTWORTHY AI PORTFOLIO
• Available for public access and
improved understanding
AIS ETHICS IN ACTION SINCE 2015
https://engagestandards.ieee.org/ieee • Enhances confidence in public and private entities that wish to realize the
benefits of AI ethics certification in the absence of or as a complement to
certifaied.html broadly accepted and enforced regulations for AI, while mitigating risks,
liabilities and adverse impacts on their reputation and market share.
© Copyrighted to IEEE
• The demand for trustworthy systems will continue to grow for the foreseeable
future. Addressing Trustworthy AI system implementations continues to be of
interest for all market stakeholders
• The IEEE SA’s adaptative and customizable approach to support your digital
transformation journey with its standards, training and CertifAIEd adaptive risk
certification methodology offers a path forward
© Copyrighted to IEEE
Thank You
twitter.com/ieeesa
certifaied@ieee.org or AIE-GET@ieee.org