Ageism, AI, and the Misunderstood Progress of DEIA

A cartoon doing the rounds on LinkedIn this week captures a truth many of us recognise: ageism in recruitment is real, and it’s getting worse. But while we must confront it head‑on, we shouldn’t let the lack of progress on age discrimination overshadow the genuine strides made in inclusivity for other groups. The DEIA movement has achieved remarkable progress in areas such as gender, ethnicity, disability, and neurodiversity — and those gains matter. They show what’s possible when awareness meets accountability.

So why has ageism become so acute, particularly in recruitment? In my view, the answer lies partly in AI‑driven hiring systems — whether embedded in Applicant Tracking Systems (ATS) or deployed through external platforms. Many of these models originate in the US, where asking for “years of experience” is routine and lawful. In the UK and Europe, it isn’t. Yet the algorithms built on US data often carry those assumptions into markets where they simply don’t belong.

AI itself isn’t biased. But the Large Language Models (LLMs) that underpin many recruitment tools can be — because they learn from human language, and human language reflects human bias. When an ATS associates words like dynamic or high‑energy with “ideal candidates”, it’s not assessing capability; it’s mirroring cultural patterns that equate youth with potential. In reality, those descriptors often mean resourceful, thorough, or experienced — qualities that have nothing to do with age.

The problem is compounded by the fact that many organisations deploy these systems without understanding how they actually work. They assume automation equals objectivity. It doesn’t. AI will always reflect the data and design choices behind it — and in agentic systems, there’s a strong positivity bias: the model seeks to reinforce what the user signals as “good”. If the user’s unconscious bias favours younger candidates, the system will amplify it.

It’s only a matter of time before lawsuits start landing — particularly for UK and European companies using US‑built recruitment products in jurisdictions they weren’t designed for. But we don’t have to wait for litigation to act.

Awareness is part of the solution. For jobseekers, that means tailoring CVs, tone of voice, and language to align with how AI interprets content — without compromising authenticity. For those of us involved in hiring, it means consciously checking our own bias, questioning the tools we use, and supporting older colleagues in their job search wherever we can.

Ageism is discrimination, not efficiency. And while AI may have accelerated the problem, it can also help us fix it — if we design, deploy, and govern it responsibly.