↑Part of: Sam Altman rejects slowdown calls, says AI industry can police safety

OpenAI and Anthropic call for AI slowdown after researcher resignation and apocalyptic warnings

Topic: technologyRegion: asia pacificUpdated: i2 outletsSources: 5Spectrum: Mixed⏱ 3 min read

📰 Scored from 2 outletsacross 1 Left 1 Center How we score bias →

Story Summary

SITUATION

After Anthropic researcher Jacob Poxon resigned and Anthropic colleague Evan Hurbinger warned of a greater than 10 per cent chance that AI would 'kill all humans' within the next decade, OpenAI and Anthropic leaders joined Elon Musk in publicly calling for a slowdown in AI development (per smh.com.au). Outlets differ on emphasis: the Sydney Morning Herald highlights internal resignations and explicit apocalyptic odds, while the aggregated American Bazaar item frames the move as leadership-level consensus without the same focus on individual warnings (per news.google.com, smh.com.au).

Coveragetap to expand ▾

Spectrum: Mixed🌍Asia: 1 · Other: 1

Political Spectrum

Position is inferred from coverage mix.

i2 outlets · Center

Left

Center

Right

Left: 1

Center: 1

Right: 0

Geography Coverage

Distribution of where coverage is coming from.

i2 unique outlets · Dominant: Asia

All2 Asia1 · 50%Global1 · 50%

KEY FACTS

Anthropic researcher Jacob Poxon resigned and accused companies of 'gambling with our lives' (per smh.com.au).
Sam Altman of OpenAI and Dario Amodei of Anthropic, joined by Elon Musk, publicly called for a slowdown on artificial intelligence development (per smh.com.au).
The American Bazaar aggregation reported that OpenAI, Anthropic and SpaceXAI leaders backed calls to slow AI development, indicating broader industry alignment in other outlets (per news.google.com).

HISTORICAL CONTEXT

The immediate backdrop is the wider global instability after the March 2026 coordinated U.S.–Israeli strikes on Iranian infrastructure and air defenses, a campaign that has produced a cycle of Iranian military responses across 2026 and shifted geopolitical attention and risk perceptions worldwide.

Structurally, current debates sit atop regulatory frameworks such as the EU’s General Data Protection Regulation (effective May 25, 2018) and the October 2023 U.S. executive actions on artificial intelligence that pushed agencies to issue safety and transparency guidelines.

Brief

OpenAI and Anthropic’s leaders publicly backed a call to slow frontier AI development after a string of researcher resignations and apocalyptic warnings inside the labs.

Anthropic researcher Jacob Poxon resigned, accusing the companies of 'gambling with our lives', and his former colleague Evan Hurbinger said there is a greater than 10 per cent chance that AI would 'kill all humans' within the next decade; those statements prompted Sam Altman (OpenAI) and Dario Amodei (Anthropic), joined by Elon Musk, to urge a slowdown (per smh.com.au).

The Sydney Morning Herald foregrounds the internal turmoil — naming resignations and named probability estimates — whereas the American Bazaar aggregation cited by Google frames the move as leaders across OpenAI, Anthropic and SpaceXAI aligning on a pause without the same focus on individual researchers' stark odds (per news.google.com, smh.com.au).

Both accounts confirm a rare public unity among top AI executives calling for restraint, but they differ on whether the story’s core is leadership consensus or alarm from within research ranks.

Dario Amodei is quoted in the Sydney Morning Herald warning of the risks of unfettered AI growth and recalling his May 2025 prediction that half of entry-level white-collar jobs could be obliterated within five years; that earlier forecast helps explain why corporate leaders are taking internal warnings seriously now (per smh.com.au).

Reporting so far distinguishes confirmed events — Poxon’s resignation, Hurbinger’s public probability claim, and Altman/Amodei/Musk’s joint call for slowdown — from wider claims about exact timelines for existential risk or precise economic outcomes, which remain contested and unquantified beyond the cited remarks (per smh.com.au, news.google.com).

Going forward, observers will watch whether other leading labs endorse concrete safety measures or binding moratoria and whether regulators or investors press for enforceable limits; sources so far document rhetoric and resignations, not new binding industry-wide rules (per smh.com.au, news.google.com).

Why it matters

Workers in entry-level white-collar roles face documented disruption: Dario Amodei previously predicted half of those jobs could be 'obliterated' in five years, a concrete economic risk named in source reporting (per smh.com.au). 2) Researchers inside Anthropic have asserted immediate safety risks—Jacob Poxon’s resignation and Evan Hurbinger’s greater-than-10%-chance claim show that some employees judge development practices as posing existential hazards (per smh.com.au). 3) Corporate leaders and investors stand to gain from slower deployment: public calls by Sam Altman, Dario Amodei and Elon Musk could protect market valuations and reputations while buying time to structure governance (per smh.com.au, news.google.com).

What to watch next

Whether OpenAI, Anthropic or SpaceXAI publish or endorse a specific, time-bound pause or moratorium on frontier model training within the next 30 days (per smh.com.au, news.google.com). 2) Whether additional researchers at Anthropic or OpenAI resign or publish whistleblower accounts following Jacob Poxon’s departure (per smh.com.au). 3) Whether Dario Amodei or Sam Altman propose concrete safety standards or verification mechanisms to regulators by year-end (per smh.com.au). 4) Whether investors or boards at leading labs demand operational changes or halt funding tied to deployment thresholds referenced by company leaders (per smh.com.au, news.google.com).

Where sources differ

7 dimensions

Framing differences

smh.com.au frames the story around internal resignations and explicit apocalyptic warnings from named researchers; news.google.com’s aggregation frames it as leadership-level consensus across OpenAI, Anthropic and SpaceXAI without the same emphasis on individual researchers (per smh.com.au, news.google.com).

Disputed or unclear

No source provides independent verification for Evan Hurbinger’s probabilistic claim beyond the quote; the catastrophic-probability assertion remains unverified (per smh.com.au).
No source confirms that any industry-wide binding moratorium or regulatory action has been adopted following the calls to slow development (per smh.com.au, news.google.com).

Omitted context

No source in this pack details any regulator actions, legal constraints, or ongoing government inquiries that preceded or followed the slowdown calls; regulatory responses are not mentioned (per smh.com.au, news.google.com).
No source provides independent data on recent model training runs, compute thresholds, or specific technical practices that researchers say prompted resignations (per smh.com.au, news.google.com).
No source mentions measured civilian or economic impact statistics tied to AI deployment beyond Amodei’s May 2025 job-displacement prediction (per smh.com.au, news.google.com).
No source cites investor holdings, executive compensation, or corporate governance documents that would clarify financial incentives behind public calls for pauses (per smh.com.au, news.google.com).

Conflicting figures

smh.com.au reports Evan Hurbinger’s 'greater than 10 per cent' chance that AI would 'kill all humans' within the next decade; no numeric probability appears in the news.google.com excerpt (per smh.com.au, news.google.com).

Disputed causality

smh.com.au presents a causal chain: Jacob Poxon’s resignation and Hurbinger’s warning prompted Altman, Amodei and Musk to call for a slowdown; the aggregation in news.google.com reports the leaders’ alignment but does not foreground the resignation as the trigger (per smh.com.au, news.google.com).

Attribution disputes

smh.com.au attributes the resignation, the quote about existential risk, and the leaders’ joint call to named individuals; news.google.com attributes leadership alignment to OpenAI, Anthropic and SpaceXAI but omits named researcher quotes (per smh.com.au, news.google.com).

Sources

5 of 5 linked articles