MIT Technology Review asks whether AI could wipe out humanity, but the debate is subscribers-only

What could be downloaded from this link is not a full article but the 'Roundtables' events page of MIT Technology Review, a monthly series of video conversations between the publication's journalists and its subscribers/MIT alumni. The session that gives this item its title, dated September 15, 2026 and already listed among past sessions, is called 'Could AI really kill us all?'.

The page itself sums up the premise of the discussion in a couple of sentences: employees at the world's leading AI labs claim there is a real chance that advanced AI systems could come to pose an extinction threat to humanity. The session promises to 'unpack' those existential fears: where they come from, whether they have any real basis or are rather alarmism, and what should be done about them if they are legitimate.

The downloaded content, however, makes explicitly clear that the recording of that conversation 'is available to subscribers only' and that logging in is required to access it. In other words, this is not an article with developed arguments, interviews or analysis, but a paid teaser: a landing page advertising the event and blocking access to the actual content behind a subscription wall. The text received contains no statement, figure, direct quote or conclusion about AI existential risk that can be reported with any confidence, beyond the generic wording of the question that gives the session its title.

The same page lists other past sessions in the Roundtables series, likewise locked behind the paywall: a discussion of the so-called 'censorship-industrial complex' in US politics, another on cellular reprogramming therapies against aging, and an earlier one on whether AI can come to 'understand' the physical world beyond language models. In every case the same pattern repeats: a brief description of the topic and a notice that the recording is accessible only to registered subscribers.

In short, it is not possible to offer here any substantive analysis of whether AI represents an extinction threat as discussed at that event, because the available material does not include the content of the conversation itself, only its promotional announcement. Anyone wanting to know the specific arguments made would need an active MIT Technology Review subscription and direct access to the recording.