These rows join FutureSearch's published…

Will OpenAI publicly release or announce an autonomous AI research intern — a system OpenAI describes as able to take on research problems on its own — before 2027-03-31?✓#1

probability: 87%

rationale:

Evidence suggests OpenAI currently possesses the internal capability to meet this milestone. In a TIME profile published 2026-08-26, Chief Scientist Jakub Pachocki stated that OpenAI has "already met its internal benchmark for an automated AI research intern" time.com. He described the upcoming Astra model as capable of implementing an experimental idea within OpenAI's codebase, running experiments, and returning results, or completing work that previously occupied a human researcher for

Weighing this narrow, near-term product milestone against other labs' much stricter quantitative thresholds for automated research affirmed our high confidence in a timely release, leaving the estimate largely unchanged.

Weighed against 12 other forecasts.

kz.linkedin.comhashe.comtechflowpost.comtimesofindia.indiatimes.comlifearchitect.ai+221 more

Will any Google DeepMind Frontier Safety Framework report or model card published before 2027-12-31 state that a machine-learning R&D alert threshold has been reached?✓#2

probability: 30%

rationale:

Status Quo and the Benchmark Gap The final estimate is 30%. As of August 2026, no Google DeepMind (GDM) report or model card has declared an ML R&D alert threshold crossed. The gap between current capabilities and the threshold remains stark. GDM's Frontier Safety Framework (FSF) v3.1 defines a "rule-out" threshold for the Critical Capability Level (CCL) at 90% pass@1 on its internal research-engineering benchmark (GRB), with the alert threshold set "marginally earlier" storage.googleapis.com. The August

Assessing the potential for softer observational triggers and the rapid progress of narrow AI research tools at peer labs slightly raised our estimate that a formal alert threshold will be crossed despite the wide gap in rigorous benchmark performance.

Weighed against 11 other forecasts.

deepmind.googlelatimes.comchatforest.comfrontierriskmonitor.orgaireiter.com+159 more

Will a binding legal restriction on the use of AI systems to conduct AI research be in force in either the United States or the European Union before Anthropic prices its initial public offering?✓#3

probability: 2%

rationale:

Short Time Window

Evaluating this timeline alongside longer-term estimates for AI governance regimes confirms that the window for any new legislation to be drafted, passed, and enacted is extremely narrow. Current expectations point to Anthropic pricing its initial public offering around October 2026, with a projected median date of 26 October 2026 futuresearch.ai. Even if the IPO timeline slips into early 2027 futuresearch.ai, there is very little time for an entirely novel category of binding l

Evaluating this timeline alongside longer-term estimates for AI governance regimes confirmed that the legislative window is too tight for novel restrictions to pass before the expected IPO, slightly depressing the estimate to the baseline tail risk of an emergency measure.

Weighed against 9 other forecasts.

cliffordchance.cometcjournal.comaugment.marketdebateus.orgeuaiactguide.com+423 more

Ask a followup

Ask a followup