Why are smart companies slowing down hiring in the AI era?
Quick Answer: Companies are reducing hiring because AI-augmented workers deliver 30% more output, but the real issue is convergence — every candidate’s portfolio looks identical when AI does the production. The smartest organisations now hire for judgement, taste, and critical thinking rather than execution speed. Goldman Sachs reports 34% of organisations cut hiring due to AI productivity gains. The differentiator is knowing when not to use AI.
- Goldman Sachs data: 34% of organisations cut headcount due to AI, and adoption is accelerating
- The convergence problem: when everyone uses the same AI tools, portfolios become indistinguishable — Google, Anthropic, McKinsey are going the opposite direction
- “Extra 30%” means strategic thinking, original research, cross-domain pattern recognition, and building trust — not just faster output
- The US justice system’s COMPAS algorithm shows AI-assisted decisions can embed systemic bias — humans rubber-stamped discriminatory outputs
- The shift that’s coming: hiring managers will ask “Where do you not let AI near your process?” — the pause is where value lives
The US justice system deployed COMPAS, an algorithm to predict recidivism. It predicted Black defendants would reoffend at nearly twice the rate of white defendants with similar histories. Judges trusted the system and rubber-stamped its outputs. The same pattern is now emerging in hiring: when humans defer to AI-generated assessments without critical evaluation, systemic biases get encoded and amplified at scale.
Frequently Asked Questions
How is AI changing hiring in design and knowledge work?
AI-augmented workers produce roughly 30% more output, leading 34% of organisations to reduce headcount (Goldman Sachs, 2025). But the deeper shift is qualitative: when every candidate uses the same AI tools, portfolios converge. Companies like Google and Anthropic are counter-intuitively investing in human research and interviewing because they understand that AI amplifies but doesn’t replace original thinking, judgement, and taste.
What is the AI convergence problem in hiring?
The convergence problem occurs when knowledge workers all use the same AI tools — their work starts looking identical. Portfolios, case studies, and even interview responses become indistinguishable. Companies are discovering that the real value lies not in AI-assisted production but in the human capabilities AI cannot replicate: critical thinking, cross-domain pattern recognition, ethical judgement, and knowing when NOT to use AI.
What skills should designers develop to stay hireable in the AI era?
Five capabilities AI cannot replicate: the ability to challenge a brief rather than just execute it, knowing when to take a position even when the data is ambiguous, reading a room in ways that data cannot capture, making ethical calls in grey areas where there’s no clear answer, and building trust through authentic human connection. These skills improve with deliberate practice, not AI acceleration.
What can the justice system teach us about AI-assisted decision making?
The US justice system’s COMPAS algorithm — designed to predict recidivism — was found to predict Black defendants would reoffend at nearly twice the rate of white defendants with similar histories. Judges, trusting the algorithm, rubber-stamped its outputs. The lesson: when humans defer to AI without critical evaluation, systemic biases get encoded and amplified. This pattern is now repeating in hiring, performance reviews, and portfolio assessment.