WorkSeptember 16, 2026via Financial Times Technology

AI can forecast the future. Should we let it?

Why it matters

As AI systems demonstrate forecasting parity with elite human predictors, enterprises and policymakers face a novel decision: adopt AI forecasting for high-stakes predictions (markets, policy, resource allocation) or treat it as advisory only. The capability is real; the governance and liability questions are not.

Key signals

  • AI forecasters' performance indistinguishable from superforecasters
  • Caveats noted but not specified in headline
  • Implication: forecasting as a profession/decision-making process is changing
  • Published September 2026—timeframe suggests this is a maturing capability, not a breakthrough
  • FT framing suggests policy/societal debate, not pure technical achievement
  • AI forecasters perform at parity with human superforecasters
  • Raises deployment and ethics questions for workplace decision-making
  • Framed as commentary on trust and governance, not a technical breakthrough

The hook

AI forecasters now match human superforecasters—but the real question is whether we should trust them with prediction markets, policy, and money.

The skill of specialised AI forecasters seems indistinguishable from much-vaunted superforecasters. But there are caveats

The week's key stories, every Friday.

ONE BRIEFING · EVERY FRIDAY · FREE

Free. Unsubscribe anytime.

AI can forecast the future. Should we let it? | KeyNews.AI