Executive leaders should create a culture of
responsibility backed by resources that enable
responsible AI from design to end-of-product use
and beyond. These steps are recommended:
1. Know the legal duties and regulatory
constraints:
Leverage existing guidance, such as the Institute
of Electrical and Electronics Engineers’ (IEEE)
Code of Ethics,
1
UNICEF’s Policy Guidance on
AI for Children
2
and World Economic Forum
guidance,
3
as well as the guidance contained
in this toolkit and guidelines for the product
team, AI labelling system, and resources
for parents and guardians and children and
youth. Commit to internal and, if possible,
external AI oversight. Report compliance and
leadership measures publicly and in simple
language so buyers can understand.
2. Build a diverse and capable team:
Include ethicists, researchers, privacy
specialists, educators, child development
experts, psychologists, user-experience (UX)
designers and data scientists. Collaborate with
non-profit organizations and educational and
research institutions for more expertise.
3. Train your team and provide resources for
success with this checklist:
Educate team members about the importance
of responsible and trustworthy AI and provide
them access to the skills, tools and time
they need to execute your vision. Have open
dialogue about unintended consequences,
possible worst-case scenarios, and the reasons
for ensuring your teams are considering the five
AI characteristics critical to putting children and
youth FIRST (Figure 2).
For more information, refer to the product team
guidelines, which offers detailed guidance on
the five areas.
4. Offer expertise to inform development of
regulations, standards and guidance:
Contribute to public forums on how AI is
being used in your products or services.
Share your experience in proposing guidance
and requirements.
5. Welcome principled efforts to label
products and services:
These should be done according to the potential
impact of AI on users. Endorse and participate in
activities to develop labelling and rating standards.
Label your offerings to help consumers make
informed choices based on recommendations
about, for example, user age, accessibility factors
and whether a camera and microphone are being
used. For additional information about labelling
recommendations, see the AI labelling system.
Actions
Putting children and youth FIRST checklistFIGURE 2
Company culture and processes address ethics and bias concerns regarding
how AI models are developed by people and the impact of AI models in use.
AI models interact equitably with users from different cultures and with different
abilities; product testing includes diverse users.
Offerings reflect the latest learning science to enable healthy cognitive, social,
emotional and/or physical development.
The technology protects and secures user and purchaser data, and the
company discloses how it collects and uses data and protects data privacy;
users may opt out at any time and have their data removed or erased.
The company explains in non-technical terms to buyers and users why AI is
used, how it works and how its decisions can be explained. The company also
admits AI’s limitations and potential risks and welcomes oversight and audits.
Source: World Economic Forum
Artificial Intelligence for Children 8