ICTworks

The United Nations Convention on the Rights of the Child promotes the right of young children and adolescents to participate in decision-making in policies, processes and practices that affect their lives.

UNICEF is developing a children’s rights perspective in the rapidly-emerging sphere of artificial intelligence. Adolescent Perspectives on Artificial Intelligence reports on workshops conducted in 2020 with 245 adolescents in five countries. It includes:

• Adolescents’ views and aspirations on AI;
• Nature of their experiences with AI;
• Their opinions on the risks and opportunities of AI.

 Youth Knowledge of Artificial Intelligence

When asked what comes to mind when they hear the term “artificial intelligence”, most adolescent participants across all workshops referred to science fiction. The Terminator, the Matrix, the Iron Man, Robots, Mark 50, and Ultron were popular examples.

They also mentioned voice assistants (Siri, Alexa, Google Assistant). Brazilian participants referred to artificially intelligent banking applications such as BIA, introduced by Bradesco, a Brazilian bank. They also mentioned Aurea, a shop assistant app linked to a retail company in Brazil called Magazine Luiza.

At one extreme, some adolescents in Manaus, Brazil had never heard of AI. However, the participants in New York were exposed to many technologies underpinned by AI yet could not always differentiate between technologies that are powered by AI and those that are not.

In South Africa, most of the participants described AI as computers that could do things that people do. At the same time, some appeared to have a limited understanding as they referred to examples that did not qualify as AI

According to participants in São Paulo, AI could augment and enhance human capabilities. They also referred to humans as flawed beings, susceptible to racism, and said that AI could enable more positive human attributes.

A more sophisticated AI concept was revealed through references to AI functionality and capability compared with human capability. In Chile, most participants agreed that AI requires a lot of data to work properly, while a few were aware that some AI systems can work with small data sets, citing as examples “small neural networks” and “simple motion sensors”.

Adolescent Views on Artificial Intelligence

1. Decision-making about AI is adult-centric.

In a few of the workshops, adolescents said that they were often consulted but were not considered to be decision-makers. For these participants, AI systems like many others, are adult-centric.

This view was specifically highlighted in Chile and São Paulo, Brazil, where participants emphasised the need for adults to acknowledge that adolescents and children have independent insights, experiences and views and that they are capable of making decisions.

2. Learning about AI in homes and among peers

Most participants revealed high personal engagement with AI in their homes and among their peers. Most also said that they do not learn about AI at school, as part of the curriculum. A few exceptions include two participants from São Paulo who had taken elective courses at school, while in Chile, some had designed AI applications at school.

3. Low awareness of AI risks

While some were able to highlight the risks associated with AI and the need to exercise caution, most either took risks for granted or considered them matter-of- factly, displaying low risk-awareness of AI.

4. Critical views about AI

While many reflected low levels of critical awareness of AI, some voiced strongly-held critical views about the social inequalities produced by AI, and the opportunities, risks and threats it can present.

5. Sophisticated and complex understandings

In each of the workshops, a few participants had limited exposure to AI technologies, and an elementary understanding of AI. Many however, reflected on their regular engagement with AI systems and some of these participants revealed complex and sophisticated understandings of AI in ways that were unexpected.

6. Parents are crucial to mitigating risks

Across all workshops, participants thought that parents have to be made aware of AI and can play a crucial, informed role in mitigating the risks for their children. While some suggested that parental controls over children’s engagement with AI should be allowed, others thought that parents could also invade their privacy. Many felt that parents could help to empower children in their use of AI.

7. Agency to confront AI risks and opportunities

While in some cases, participants felt that software engineers and parents can do more to manage risks posed by AI, many also mentioned actions that they could take as individuals to mitigate risks. Others also considered that collective youth action could involve engaging companies and governments to address risks posed by AI.

8. More AI education needed

Not all adolescent participants were familiar with digital technologies and AI systems. Some had never heard of AI. However, they were able to participate meaningfully in discussions on AI and its current and future implications for adolescents.