The class of ChatGPT is graduating: Are UK employers ready? 

This year’s graduates are the first to enter the workforce with generative AI as a consistent part of their education. For many, free tools like ChatGPT are central to how they research, write, and solve problems. At the same time, employer expectations have shifted, with AI skills becoming a common requirement in early-career roles. The question isn’t whether graduates are ready for work, but whether employers are equipped to guide them.

The disconnect in AI expectations

Employers are raising their expectations around AI at a time when graduates are entering the toughest job market since 2018. As AI tools become more embedded in day-to-day work, employers place greater value on candidates who can use them critically, responsibly, and transparently from day one. 

Gaps within organisations themselves compound this pressure. Research shows that 39% of employers use conversational AI, yet 61% admit they have no staff directly working with it, and only 11% offered AI-related training last year. This creates a clear tension: businesses want graduates who can use AI responsibly while still defining what responsible AI use looks like internally.  

The same gap appears in education. A recent study found that 41% of UK degree-awarding institutions have no publicly accessible AI policy. This lack of clarity makes it harder for students to understand when AI use is appropriate, how it should be disclosed, and what responsible use actually means. 

Without formal guidance, learning often stays informal. Seven in 10 alumni report teaching themselves AI through experimenting, while only a small minority have completed a formal university course focused on AI. That distinction matters. Being comfortable with AI is not the same as knowing how to use it well. Many students know how to generate outputs, but they may lack the judgement to interrogate or verify them. That’s a gap businesses need to address during the transition from education to employment..

Rethinking onboarding for recent graduates 

Bridging this gap requires a shift in how businesses onboard new hires. AI is changing the nature of work itself, shifting the value of human contribution toward judgment, creativity, and decision-making. Knowing how to prompt a tool is a starting point, but the real differentiator is the ability to apply AI to solve complex problems, question automated outputs, and explain decisions clearly.

That means employers need to move beyond the assumption that younger hires already “know AI”. Comfort with a tool doesn’t automatically mean someone understands its risks, limitations, or appropriate business use. Research found that a large proportion of organisations rely on on-the-job training rather than structured education and training programmes, while only 13% of graduate schemes include AI training. Without clearer onboarding, businesses risk leaving new starters to figure it out for themselves. 

Effective onboarding should do more than introduce tools. It should include: 

• Clear guidelines on responsible AI use: when tools can be used, where restrictions apply, and what accountability looks like. 
• Training on risk and limitations: including data privacy, hallucinations, bias, and reliability.
• Role-specific examples: showing how AI can (and shouldn’t) be used in day-to-day work. 
• Opportunities to experiment safely: so employees can test approaches and build confidence without risk. 

This matters because the biggest risk is usually misunderstanding rather than deliberate misuse. Graduates may be comfortable with AI-generated outputs but less aware of hallucinations, bias or the privacy implications of inputting sensitive information into third-party systems. They may know how to prompt, but not how to verify. That is why onboarding should go beyond basic tool usage and focus on the skills that matter most in practice: how to sense-check outputs, how to identify unreliable information, how to handle data safely, and how to explain decisions clearly. 

Final thoughts 

The transition to an AI-enabled workforce is not the responsibility of one group alone. Universities must continue to evolve how they teach AI, embedding it into learning in a way that builds both capability and critical thinking. At the same time, businesses should recognise that graduates are arriving with potential, not perfection. 

Familiarity with AI is a starting point, but its value depends on whether employers create the structure, confidence and clarity needed to turn that familiarity into effective workplace practice.

‍