Skills for a world
with AI

One of the main characteristics of an AI-rich world is the automation of both complex and routine tasks. AI systems can now generate text, images, audio and video, analyse patterns and support decision making. They can perform in ways that were previously understood as distinctly human.

What does this mean for learners?

AI can now replicate many key skills including generating, retrieving, analysing, evaluating and synthesising information, which are traditionally learned in schools.

The OECD and Lodge and Loble (2026) point out that some beneficial cognitive offloading can occur when we use AI. This is when our lower-level cognitive load is reduced, for example, experienced authors using AI to proof-read their work. This can free up limited working memory space to focus on essential, higher-level cognitive tasks. In this way, AI has the potential to enhance learning and productivity, but only if it takes on the tasks that learners have already mastered.

Moreover, educationalists including Oakley et al. (2025) and Lodge and Loble (2026), warn that a major risk of AI is what is known as detrimental, or harmful, cognitive offloading. In traditional education, learning helps construct ‘schemas’ or knowledge structures in our long-term memory. We then use these schemas to acquire even more knowledge and also skills. Detrimental cognitive offloading occurs when we bypass this process, instead relying on AI systems to do the work for us. As a consequence, we are not developing our own knowledge and complex thinking skills, or the resilience to do some tasks that we might prefer not to do.

We also miss the chance to build and strengthen the sustained attention needed to focus on a task that underpins these skills. Allowing machines to think for us can therefore undermine the cognitive capabilities that education aims to develop. The result is a tension between short-term efficiency and long‑term intellectual growth.

A recent joint study conducted by Cambridge researchers with Microsoft Research supports this view. It concluded that large language models (LLMs) such as ChatGPT or Microsoft Copilot can help students clarify, explore and contextualise learning material. However, traditional note-taking remains critical for reading comprehension and knowledge retention. The study forms part of a growing and rapidly evolving body of research evidence that is continually shaping our understanding of AI’s role within education.

What does this mean for learners?

To preserve their long-term intellectual growth, learners need to develop strong metacognition, human judgement, personal strategies and self-discipline around:

• when to use AI
• when not to use AI
• actively making sense of rather than passively consuming AI outputs.

A defining feature of AI‑integrated societies today is a vastly different information ecosystem. Eliot Higgins, the founder of Bellingcat, argues that in the second half of the 20th century, knowledge creation and information sharing largely came via institutionally mediated systems. This means that previously, institutions such as governments, the media and education systems filtered and verified information. However, we now share information directly with one another via the internet and social media. Information circulates at speed and at ever-increasing volume, often without having been verified by those with expertise or authority. What’s more, information is increasingly organised by algorithms. As a result, we play a greater role in interpreting and distributing raw information, and in creating knowledge.

In this context, every act of viewing, liking or sharing online content, both AI-generated and human-generated, becomes part of a wider epistemic (knowledge-related) process. The risk is not simply that it exposes us to false or misleading information. It is also that we develop multiple different, incompatible ‘realities’, seeing or hearing information that aligns with our existing beliefs and sharing it within like‑minded communities.

What does this mean for learners?

Higgins emphasises that education needs to equip young people with the thinking skills they need to be positive participants in this new information system. This includes:

• understanding how knowledge is constructed
• questioning sources and intent
• reflecting on their own assumptions
• recognising their responsibility to others when sharing information.

Without this, an AI‑rich information environment can undermine robust deliberation – the ability to evaluate options and challenge assumptions.

Although knowledge is not the focus of this guide, it is worth mentioning that we also need to consider the knowledge contexts in which learners develop and use the critical skills they need for an AI-changed world. Arguably, the need for teacher-led rather than student-led learning becomes more important. Relying on students’ own interests to steer the topics they explore in school could limit them. They will only see the content that the algorithm-constructed feeds on their phones promote. This could limit their opportunities to acquire social and cultural skills and to expand their horizons.

At a societal level, AI reshapes the division of labour between humans and machines. It therefore raises questions around power, responsibility, sustainability, and equity. Furthermore, AI systems are developed, deployed and governed unevenly, with the potential to increase existing social and global inequalities.

From an environmental perspective, AI brings new considerations and implications. There is hope that AI can help tackle some of the world’s environmental challenges, for example, the technology is already being used to map the destructive dredging of sand and to chart emissions of methane, a potent greenhouse gas. However, there are negative environmental impacts too from AI infrastructure, such as the necessary ongoing cooling processes at AI data centres which are highly intensive in their water use, adding to water scarcity concerns.

In this context, human agency (our ability to make independent choices), accountability and ethical judgement become more important, particularly in high‑stakes domains such as education, healthcare and public decision‑making.

What does this mean for learners?

Taken together, these trends point to a future marked by rapid change, uncertainty and continual adaptation. We all need to be capable of lifelong learning, ethical reasoning and responsible participation as humans in control of technology and its outputs. We cannot assume that technological progress alone will have positive outcomes.

Our Digital Literacy and Wellbeing curricula for Cambridge Primary and Cambridge Lower Secondary are particularly rich in skills that can be taught with direct reference to AI. For example, at Stage 7 of our Digital Literacy curriculum students ‘apply AI tools to refine and enhance digital outputs, including multimedia, while retaining the role of the main creator, and explore the benefits and limitations of using AI supported media tools in a range of contexts’.

In Stages 7 to 9 of our Wellbeing curriculum, as part of the ‘Navigating my world’ strand, students ‘recognise the importance of permissions online and the risks associated with sharing images and information with others. They are also expected to ‘explore opportunities and strategies for managing their own personal safety, and to demonstrate metacognitive strategies in their learning’.

Find out more about Cambridge Primary and Cambridge Lower Secondary Digital Literacy.

At Cambridge Upper Secondary and Cambridge Advanced stages, learners have the opportunity to develop skills in AI-specific contexts as well as build on their transversal skills.

For example, in Cambridge International AS & A Level Computer Science, learners explore the impact of AI, including social, economic and environmental issues. This helps them understand both the opportunities and the risks of AI.

Other subjects develop AI-ready skills in wider contexts. For example, Cambridge Global Perspectives encourages learners to research, analyse evidence, understand different viewpoints and work together on projects. This is important in a world of AI-generated content and technologies, for learners to compare information from different sources or discuss how technology affects society.

The Cambridge Pathway offers clear progression in primary and lower secondary skills development, through Cambridge IGCSE and O Level and into advanced study and beyond.

At earlier stages, students develop and apply AI-relevant skills by engaging with information that is typically more familiar, more concrete (rather than abstract), and lower in volume and complexity. For example, in our Computing curriculum, students ‘explore computational thinking, including algorithmic decision-making processes in familiar contexts’. These contexts are much simpler than those covered at Cambridge International AS & A Level Computer Science.

Similarly, systems thinking is a relatively advanced concept that is useful for understanding the connections between, and implications of interactions within and between, AI technologies and humans. The foundations of systems thinking are present in Cambridge Lower Secondary Science. Students study the connections within ecosystems in simplified forms, providing a foundation for more complex applications later, such as those in Cambridge International AS & A Level Marine Science.