The BBC Computer Literacy Project for the AI age
I have spent the better part of four decades in technology, from punch cards and ICL mainframes to the modern AI stack. One moment stands out as particularly relevant: the BBC Computer Literacy Project in the early 1980s.
That project was not about training people to be hobbyists. It was about making sure the country understood the coming shift before it arrived. The lesson still matters.
We are now in the middle of another foundational shift. The issue is not whether people can use AI. It is whether they understand what it is, what it is not, and what it means for work, identity, and trust.
We do not need a few more prompt engineers. We need a basic public literacy around AI, sovereignty, and reasoning.
1. We still need to teach people how to state intent clearly
One of the biggest mistakes in the current AI boom is that people treat it like a magic box that will turn a vague idea into a good answer. It will not.
AI is very good at taking structure and turning it into output. It is much worse when the input is muddled, vague, or contradictory. That means society needs a better understanding of how to describe intent, constraints, and acceptable answers.
That is not a technical skill only. It is a literacy skill.
2. Identity is being redefined in public
As routine work gets automated, people are being asked to rethink what their value is. That is uncomfortable. It is also necessary.
Human value is not just in doing repetitive work. It is in judgement, context, accountability, and the ability to reason under ambiguity. The public conversation has been too focused on replacement and too little on what kind of work people should move toward.
That is why literacy matters. Without it, the question becomes “what do I do now?” instead of “what am I actually good at, and how do I move up the stack?”
3. People need a hierarchy of truth
The dangerous part of AI is not only hallucination. It is the tendency to confuse a probabilistic system with a deterministic truth engine.
People need to understand the difference between raw model output, a cloud service, a local governed system, and a fact with evidence behind it. They need to understand provenance. They need to understand why a locally governed system can be more trustworthy for some tasks than a broad public model with no traceable chain of reasoning.
The new curriculum
If I were designing a modern public AI literacy programme, I would focus on three things. First, sovereign parity: understanding why local compute and local control matter for privacy, governance, and operational resilience. Second, agentic logic: understanding how to orchestrate digital work rather than simply chatting with a bot. Third, algorithmic governance: understanding bias, risk, provenance, and decision accountability.
That is not a luxury. It is the basic knowledge people need to survive the next decade without mistaking capability for wisdom.
In the 1980s, the BBC helped a generation understand the computer. In 2026, we need that same seriousness around AI. Not because we all need to become engineers. Because we all need to understand what kind of world we are building.