Unprioritized skills aren’t safe

If a skill still looks uniquely human, that may mean AI companies simply haven’t chosen to target it yet — not that it’s in principle hard to automate. As firms change priorities, formerly 'safe' academic and professional skills can become automation targets, so planning should assume eventual capability rather than perpetual immunity. — Shifts the debate about automation from 'can AI do X?' to 'will AI firms prioritize X?', affecting education, labor policy, and institutional preparedness.