Trapped Workers: Which Moves Actually Change a Worker's Options? • Bipartisan Policy Center

Oh my goodness, buckle up, because the Bipartisan Policy Center has just handed us the most *selective* four-year trend imaginable — and honestly, who among us wouldn't choose to slow down a little! ✨ Net inflow into more AI-exposed work peaked in 2020 at 2.2 million weighted transitions and gently settled at 0.2 million by 2024. That is not a collapse, that is a workforce exercising exquisite timing: 2020 was a busy year for absolutely everyone, and by 2024 people had simply learned the word 'no.' What magnificent restraint! Weighted transitions, no less — even the unit of measurement is giving it everything.

And the clustering news is even lovelier! 41% of workers leaving highly AI-exposed occupations land in the *very same occupational cluster*, which we are choosing to read as extraordinary loyalty — nobody wants to abandon their people mid-pivot! The 59% who do cross clusters are rewarded with reachable next jobs carrying 6.5 percentage points less AI exposure on average, and that is not a downgrade, it is a *gentle decompression*. Six and a half points! Imagine the relief of that much less exposure, like stepping out of one sauna and into a marginally different sauna. The brief calls these 'reachable next jobs'; we call them *destinations with a view*. ✨

And that title! 'Which Moves Actually Change a Worker's Options?' — a question, which means options exist, which means *choice*, the most beautiful word in economics. ✨ The Bipartisan Policy Center is doing the patient, bipartisan work of mapping the corridors, and a corridor is fundamentally a route, and a route goes somewhere or it wouldn't be a route. Nobody is trapped here — they are *contained*, which is basically 'trapped' with better branding and two-party sponsorship.