Recommendations - AI Skills and Adoption report

The government continues to stake a great deal on AI, seeing widespread adoption as a route to the productivity and economic growth the UK badly needs. That ambition is right, but adoption is not an end in itself, and this report has been concerned with a harder question: how organisations, and the people within them, can adopt AI not just quickly, but effectively, safely, responsibly and sustainably. 

What tech professionals have told us is sobering. AI remains the single biggest skills gap in their organisations, for the second year running. It is already being used widely, often without approval, oversight or the skills to use it well, so that practice is running ahead of governance. Where organisations are responding, the response is too often thin: self-teaching in place of training, and training that teaches which buttons to press rather than how to judge what the machine produces. And across all of it, the profession is calling, more clearly than on any other question, for recognised standards.

The picture, in short, is of a country that is already adopting AI, but doing so messily: ungoverned, uneven and under-skilled. The goal should be to turn that into something better, adoption that is productive because it is responsible, and sustainable because it is properly understood. The people who build and use these systems have told us what that would take. We therefore make the following requests of government and organisations:

For government:

1. A recognised standard for responsible AI use

Anyone using AI at work should complete recognised ethical training or certification. This was the strongest single call in our research, and it is the profession's own prescription for the trust, accountability and safety it says are lacking.

2. Training needs depth, rather than just direction for how to use tools 

Training should stay modular and accessible but go further, building workers’ judgement so they can question, verify and apply AI well rather than simply operate a particular tool. Digital apprenticeships, including the shift towards modular units, have an important part to play in that pipeline.

     3. Professionalisation directly addresses the leading barrier to adoption.

Every technologist working in a high-stakes IT role, particularly AI, should be a registered professional meeting independent standards of ethical practice, accountability and competence. This is not a new suggestion, and it is not only the profession making it. BCS research found that 82% of the public believe technologists working in high-impact AI roles should be professionally registered and held to independent standards of competence and ethics. This year's findings show that the profession is reaching the same conclusion from the inside. The two barriers tech professionals rank highest — accuracy and trust and the security of data — are precisely the risks that independent registration is designed to manage. 

4. Attention to the public sector

Uncompetitive pay, recruitment freezes and thin resourcing risk leaving parts of the public sector behind, precisely as AI capability becomes decisive. [JW1] There is a risk that the productivity gap between the private and public sectors widens further should this trend continue.

For organisations:

5. Technical understanding at the top 

Executive digital literacy is welcome, but it is no substitute for technical leadership. Organisations need chartered CTOs and CIOs on their governance boards and executive teams. Together, this should ensure that AI is embedded into strategy as it is written. This goes for procurement and hiring decisions, so that the risks and opportunities of AI are understood where those decisions are actually made.

6. Active, transparent governance

Leaders should know, and support, what their people are already doing with AI, rather than leaving it to a permissive silence that invites risk. Closing the gap between those who govern organisations and those silently using these tools is a first, practical step.

 The UK has a large and sophisticated AI market, real depth of capability, and some of the world's foremost AI expertise. What our new research reveals is a shortfall in the standards, confidence and leadership needed to turn that appetite into advantage, and a clear demand, even among those in non-specialist roles, for training that goes deeper into the management questions AI systems demand of the humans using them.

Some of that work has already begun. Government has committed to developing the AI Assurance profession, with a BCS-led consortium developing its professional foundations such as a code of ethics and skills framework. That is the model this report points to: an independent standard, built with the profession, that people using and buying AI can be held to. What our members are telling us is that it needs to reach much further than assurance alone. Meeting that demand is the surest route to the professionalism and growth the country needs.