Is AI Actually Going to Kill Us All?
By Brian Barrett, Zoë Schiffer, and Leah Feiger | September 10, 2026
Overview
In the 2026 conversation surrounding artificial intelligence, one question refuses to fade: could AI pose an existential threat to humanity? This week on Uncanny Valley, WIRED's podcast, the team examines several of the most consequential stories shaping the AI landscape—from a former Anthropic researcher's stark doomsday warning to Apple's latest hardware announcements, and a census report that falsely claimed Trump won the 2020 election.
A Former Anthropic Researcher's Warning
Among the top stories this week: a former researcher at Anthropic—one of the world's leading AI safety companies—has gone public with a stark warning about the risks posed by advanced AI systems. Their concerns echo a broader debate that has intensified throughout 2026, as frontier models grow more capable and the companies building them face mounting scrutiny over safety practices.
At the heart of the discussion: Are these warnings genuine existential alarms, or do they reflect internal tensions inside the labs themselves? Uncanny Valley digs into what the researcher said, why it matters, and how it fits into the wider AI safety conversation.
Apple's Latest Upgrades
Also this week: Apple held its September event, rolling out a new slate of product upgrades with AI features woven throughout. As Apple continues its late-but-deliberate push into on-device AI, the announcements offer a window into how mainstream consumer tech is now being shaped by artificial intelligence.
The team breaks down what's new, what's genuinely useful, and what's just marketing—plus how Apple's approach compares to rivals like Google and OpenAI.
The Census Report That Claimed Trump Won 2020
The episode also covers a newly surfaced census report that falsely asserts Trump won the 2020 election. The claim—widely debunked—raises fresh questions about how misinformation continues to spread through official-seeming documents and data, and what role AI tools might play in generating or amplifying it.
Why It Matters
From existential risk to everyday consumer devices to the integrity of democratic data, these stories converge on a single theme: AI is no longer a future concern. In 2026, it is actively reshaping how we work, what we trust, and what risks we are willing to accept.
Listen to the full episode of Uncanny Valley on WIRED.
via Wired AI