Skills for an AI World

In our consultations with children, they viewed AI as a tool that can be used for a wide range of tasks in education, work and life in general. Their understanding of AI can be restricted to familiarity with chatbots rather than recognition of its broader applications, such as for predictive analytics and recommendations systems.  

Children used a range of AI tools, predominantly chatbots like ChatGPT and Meta AI. They used these for a variety of activities, including to support with schoolwork, look for information, help with tasks and learn skills. Few children reported using AI for companion-like interactions. 

Older adolescents demonstrated a more sophisticated level of engagement with AI. They described using more powerful research tools and provided more precise detail about their activities with AI systems compared to younger peers.  

AI was seen as useful. Older adolescents described the value of AI to help with research to complement people’s knowledge, speed up tasks and perform some tasks to a higher standard than humans. 

At the same time, children shared significant worries about AI both today and in the future. They were concerned about AI’s privacy and safety risks, and most think AI cannot be trusted as it can provide false information. They were also worried about people becoming dependent on AI. Older adolescents expressed worry about AI creating job losses in the future. In general, children felt uncertain about future AI developments. 

• Learning to prompt AI was an important skill for children we consulted. Aside from this, while some children discussed the need for advanced digital skills (such as coding) and basic digital skills (like knowing how to type), children did not prioritize digital competencies. 

• Skills related to critical thinking and information literacy were seen as key. These involve knowing how to evaluate and verify information from various sources. 

• Children felt it is important to develop skills related to safety, data protection and privacy, especially understanding what information is and isn’t safe to share with AI. However, there was disagreement among children about what information should and should not be shared with AI systems. 

• Foundational skills of literacy and numeracy were also sought. These were mentioned mostly among younger children. 

• Children believed that there is a need to self-regulate AI use by limiting how big a role AI plays in life and by being intentional about what to use AI for. 

• Children felt it is important to learn skills themselves, not merely to complete tasks, but also to allow humans to maintain their autonomy. 

• Children wanted to invest in human relationships, as AI should not replace interactions traditionally carried out between humans. 

• Children placed importance on understanding AI, particularly knowing that AI is a non-sentient technology. However, there was little discussion about developing other knowledge about AI, like understanding its wider capabilities and limitations. 

• More importantly, children felt it is necessary to have pre-existing knowledge about a subject when using AI. This plays an important role in underpinning AI use, as it can help people to verify information and avoid over-reliance. 

Thriving in an AI-influenced world requires much more than just technical digital competencies. Those working on developing AI literacy must consider the full range of skills, knowledge and attitudes that children require for navigating an AI-influenced world. Skill initiatives must also meet children at their current level of understanding and use of AI, even while raising that level. 

While it is essential to develop skills in schools, this approach alone is insufficient. Children learn about AI from a range of sources, including extracurricular programmes, peers and family members. Activities and extracurricular programmes to develop relevant AI competencies should be developed and supported for children outside of schools. 

AI-related awareness campaigns can also help parents and caregivers develop their own understanding of AI (including its benefits, risks and limitations) so they can support children with the necessary competencies for safe and effective AI use. 

Digital infrastructure must be improved so that all children have equal access to the benefits of AI. Children are aware of financial and infrastructure barriers facing them and how these could restrict their use of AI. There must be meaningful connectivity through affordable internet access and device availability, particularly in rural and low-income settings. 

Skills initiatives and AI systems must be appropriate to existing levels of infrastructure and connectivity, cultural needs and linguistic factors. Even within this study’s sample of five countries in Africa, there was considerable variation in digital infrastructure and responsible AI index scores. This creates different priorities in terms of which skills should be prioritized at which point.