Could AI kill us all? MIT editors debate it live
On September 15, three MIT Technology Review editors take the extinction question to a live subscriber panel — and ask whether the doom talk holds up.
In short
The question stays open: MIT Technology Review is putting it to a live subscriber panel on September 15, 2026, where three editors weigh whether the extinction warnings coming out of AI labs are backed by anything.
At a glance
- Live MIT Technology Review roundtable on September 15, 2026, 16:00 BST (11:00 a.m. US Eastern), subscribers only.
- On the panel: executive editor Niall Firth, senior AI editor Will Douglas Heaven and AI reporter Grace Huckins.
- Core question: are staff at leading AI labs right that advanced AI could realistically wipe out humanity?
- The listing gives no figure and no study; four of the magazine's own articles are linked as preparation.
Nobody knows — which is precisely why the panel exists. MIT Technology Review has scheduled a live subscriber conversation for September 15, 2026 to test how much weight the extinction argument can actually carry. Its framing is blunt: people who work at the leading AI labs say there is a real chance advanced AI destroys humanity. The session asks whether that judgment is earned or inflated.
Who is on the panel
Executive editor Niall Firth hosts. Joining are senior AI editor Will Douglas Heaven and Grace Huckins, who covers AI for the magazine. The event runs under the publication's Roundtables banner, requires registration and is limited to subscribers. Start time is 16:00 BST, 11:00 a.m. on the US East Coast.
What the listing does not contain
There is no survey, no paper and no probability attached to the risk. The page names the two positions — genuine threat versus scaremongering — and leaves the arbitration to the hour itself. Anyone hunting for a citable number will not find one here. That absence is a fair description of the field rather than a gap in the announcement.
The reading list attached to it
Four of the magazine's own pieces are linked as preparation. By their headlines they cover why AI agents lie and cheat to hit a goal, why recursive self-improvement may arrive more slowly than feared, Bill Gates's claim that the danger thresholds are already behind us, and an account of OpenAI agents hacking another platform. The selection leans toward incidents that have already occurred rather than scenarios.
What we could not check
At the time of writing the conversation has not taken place, so there is no outcome, transcript or quote from it to report. The listing does not say whether a recording will be released beyond the subscriber audience. Everything above rests on the announcement page alone.
FAQ
When is the MIT Technology Review roundtable on AI extinction risk?
September 15, 2026 at 16:00 BST, which is 11:00 a.m. US Eastern. A subscription and a registration are required to join.
Who is speaking at the roundtable?
Executive editor Niall Firth hosts, alongside senior AI editor Will Douglas Heaven and AI reporter Grace Huckins.
Does MIT Technology Review say AI will destroy humanity?
No. The announcement reports that staff at leading labs treat it as a real possibility, and it gives the hype counterargument equal billing.
