The race to make AI safer
As AI leaders warn that the technology is advancing faster than our ability to control it, Donald Trump has dismissed some of those concerns as exaggerated. On FRANCE 24, AI philosophy professor Raphaël Millière said this summer's real-world AI security incident shows the risks are no longer hypothetical, while stressing that it does not mean "AI will kill us all". Instead, he argues the priority is coordinated action by companies and governments to manage increasingly real safety risks.
Places named in this report: China, France
Exact coordinates
china: 35.860, 104.200
france: 46.230, 2.210
Read it at France 24 See this on the map
What might happen next? AI-generated
These scenarios are written by an AI language model from the headline and summary above. They are not predictions from the newsroom, and they are not evidence of anything. Every one is given a deadline and checked against later coverage, and the score is published on the ledger — including the ones that miss.
-
Unresolved 35% Accelerated Regulatory Push
Following the warnings from AI experts, a significant international regulatory body announces a binding framework for AI safety, perhaps triggered by a specific real-world incident. This pressure forces both governments and tech companies to move quickly to demonstrate compliance. Trump's dismissal is sidelined by the perceived severity of the risk.
Watch for: EU Parliament passes a new AI safety directive within 7 days · G7 summit includes a specific, actionable resolution on AI risk management
-
Unresolved 30% Business Self-Regulation Dominates
AI companies take the lead, ignoring political rhetoric from leaders like Trump. Major tech firms release a unified technical white paper detailing new safety guardrails, effectively setting industry standards before governments can impose rules. This collaborative approach minimizes immediate geopolitical friction.
Watch for: OpenAI and Google release a joint technical paper on alignment protocols · A major AI vendor announces a voluntary 'safety escrow' for high-risk models
-
Unresolved 25% Political De-escalation/Dismissal
Trump's dismissal gains traction among his supporters and key political allies, successfully framing AI safety concerns as an elitist overreach. The focus shifts away from immediate, binding action, leading to a temporary cooling of high-level security debates in major political forums. This path minimizes immediate global conflict.
Watch for: A named US politician tweets a statement supporting the 'freedom to innovate' argument · The White House releases a statement downplaying the necessity of immediate AI safety treaties
-
Unresolved 10% Unforeseen 'Chaos' Trigger (Counter-intuitive)
Contrary to the idea that AI is a slow, manageable risk, a highly publicized, low-stakes but undeniably dangerous AI error occurs in a public sector application. This incident is not a catastrophe but serves as a tangible, undeniable proof point, overriding all political discourse. The focus immediately shifts to emergency international intervention.
Watch for: A news headline reports a specific, non-catastrophic but hazardous AI failure in a city's traffic control system · A national security advisory is issued by a major Western nation referencing the specific AI incident
Generated by llama
on 2026-09-13. Checked against later coverage after 2026-09-20.
See how these forecasts score.
ForecastGeo shows the headline and summary published by the newsroom and places the story on a map. The full article lives at the source.