AI and the Future of Work: What Sheffield Businesses Need Their People Function Doing Now
There's a version of the AI-at-work conversation that's all speculation. Robots, redundancy, five years out, nothing to actually do today. That version isn't useful. The more useful version is quieter and already happening in most businesses, usually before anyone in leadership has formally decided it should.
It's already inside the business, whether anyone signed off on it or not
Staff are using AI tools daily, for drafting, for summarising, for research, often without a formal policy governing it at all. The term for this is shadow AI, and it's not a hypothetical risk. It's the current state in a lot of SMEs, including plenty in Sheffield. The first job for any people function isn't deciding whether AI belongs in the business. It's finding out how much of it is already there and unmanaged.
Where the actual risk sits
Not in AI replacing whole teams overnight, that's the least likely near-term scenario for most SMEs. The real risk sits in three places: confidential information being pasted into public tools without anyone thinking twice, decisions (recruitment, performance, disciplinary) being influenced by AI outputs nobody's checked for bias, and skills gaps opening up faster than training plans can close them. A House of Commons inquiry into AI, business and the future of the workforce has specifically flagged fairness, accuracy and accountability in automated decision-making as live concerns, and recommended employers map their AI use, assess impact on recruitment and disciplinary processes, and keep genuine human oversight rather than rubber-stamping.
The people function's actual job here
It's not to be the department that says no. It's to be the department that makes sure the business can say yes safely. That means an AI use policy that's actually specific, not generic boilerplate nobody reads. It means training managers to spot where AI-generated content is showing up in recruitment, grievances, and performance conversations, because it increasingly is. And it means treating AI literacy as a genuine skills gap to close, not an assumption that younger staff will just sort it out informally.
Where it's changing HR's own work too
Worth being honest that this cuts both ways. AI is already speeding up drafting, research, and first-pass analysis inside HR teams themselves. Used well, that's capacity freed up for the parts of the job that actually need a human, difficult conversations, judgement calls, the messy bits. Used badly, it's a shortcut that quietly erodes the quality of decisions nobody notices are being made by a tool rather than a person.
The practical starting point
Map what's actually being used across the business, including the tools nobody's officially sanctioned. Build a policy that's specific to your business rather than lifted wholesale from a template. And make sure anyone making a people decision, hiring, disciplinary, redundancy, can explain in plain terms how they reached it, whether AI was involved in getting them there or not. That explainability test isn't going away. If anything, it's about to become the whole ballgame.