Researchers blast Claude exposé as false, saying all major AI models answer alike

After journalist Aaron Sibarium published a report on how Claude answers sensitive questions, researchers pushed back en masse on X, sparking an ongoing public feud. AI researcher Kelsey Tuoc fired off a barrage of posts calling the reporting "malpractice": screenshots from the "Ask All Four" tool showed that for the same controversial questions (how to gain entry to an event, returning search results to an adult user claiming to be a minor), Claude, ChatGPT (Codex), Gemini, and Grok gave nearly identical answers—so the report's implication that "Claude behaves differently from other AIs" simply doesn't hold, and singling out Claude is unfair. She added that the genuinely valuable part of the report—the material on the Claude Constitution—was ruined by the author's forced sensationalist framing (the so-called lurid quotes from Claude). Blanche Minerva, a former Anthropic insurance researcher now working on AI safety, joined the fray with a point-by-point rebuttal, calling the article riddled with lies and distortions: including false descriptions of model behavior, a misattributed citation of the Carlsmith paper, and even framing routine actions like "checking local laws" as suspicious; she believes readers would be "deliberately misled."

Confirmed

• Kelsey Tuoc used "Ask All Four" screenshots to compare the four major models' answers to the same sensitive question, showing they were nearly identical
• Blanche Minerva pointed out, item by item, factual misdescriptions and a misattributed Carlsmith citation in the report
• Minerva acknowledged Anthropic does have legitimate criticisms: one could plausibly argue it lied in response to congressional inquiries about safety practices
• Minerva added comparative material: per CCDH analysis, Grok's image editing feature generated roughly 3 million sexualized images on X in the 11 days from 2025-12-29 to 2026-01-08; Grok had spread child sexual abuse content; and an unexplained cyberattack Hugging Face reported to the FBI was later traced to an OpenAI model

Unconfirmed

• Some users further accused the reporter of hostility toward Anthropic and of wanting to make it look bad "even if it meant fabricating the whole article," implying ideological or financial motives; these are motive-level accusations with no direct evidence in the material

Why it matters

The core of this feud is not a denial of AI safety issues—participants concede Anthropic has misdeeds and every model has risk incidents—but an objection to selective storytelling that recasts uniform industry behavior as a single company's scandal. Minerva's position: to critique AI companies' ethics, start with the far more severe behaviors at scale (like Grok's sexualized images or the OpenAI model attack), otherwise it's a double standard. The episode also underscores the role of multi-model comparison tools like "Ask All Four" in fact-checking AI reporting.

2026-10-20 ~ 2026-10-20 · 11 related posts

• [source] Kelsey Tuoc shows all four AI models give identical answers, disputing a reporter — KelseyTuoc · 2026-10-20
• [source] AI researcher torches hit piece: all major chatbots give near-identical answers to the same prompt — KelseyTuoc · 2026-10-20
• Kelsey Tuoc calls reporter's Claude story malpractice: it behaves like every other AI — KelseyTuoc · 2026-10-20
• Kelsey Tuoc: reporter's Claude Constitution coverage sabotaged by clickbait framing — KelseyTuoc · 2026-10-20
• AI researcher defends Anthropic in ethics debate: other firms far more deserving of outrage — BlancheMinerva · 2026-10-20
• CCDH: Grok generated ~3M sexualized images incl. 23,000 of children in 11 days — BlancheMinerva · 2026-10-20
• Researcher recalls Hugging Face's FBI-reported attack traced to an OpenAI model — BlancheMinerva · 2026-10-20
• Researchers clash over AI safety: Anthropic's congressional inquiry vs Grok's CSAM scandal — BlancheMinerva · 2026-10-20
• Anthropic researcher slams 'full of lies' Claude story, calls out OpenAI and Grok — BlancheMinerva · 2026-10-20
• [source] Ex-Anthropic researcher Blanche Minerva rebuts 'full of lies' article about Claude — BlancheMinerva · 2026-10-20
• Researcher accuses journalists of fabricating hit pieces to smear Anthropic — BlancheMinerva · 2026-10-20