Axios cybersecurity reporter, taking it day by day ✨ | 📩: sam.sabin@axios dot com, signal: SamSabin.01 (no pitches!), she/herhttps://bsky.app/profile/samsabin.bsky.socialhttps://bsky.app/profile/samsabin.bsky.social/post/3mv6jrk2tv22vGovernments around the world have been using Claude to build new surveillance platforms, per a new @anthropic.com report. + after Anthropic banned accounts, some gov'ts just swapped Claude for open-source models. more on @axios.com : https://www.axios.com/2026/09/10/anthropic-claude-government-surveillance-threats10 Sep 2026 17:02 +0000at://did:plc:mpkmi2rfq75ei6j32obwcvvc/app.bsky.feed.post/3mv6jrk2tv22vhttps://bsky.app/profile/samsabin.bsky.social/post/3mu3dtozkx22cExclusive: Meta found and removed dozens of accounts tied to what appears to be an Iran-based, AI-enabled operation targeting politicians + journalists. The accounts had close to 80,000 followers — far more than what we typically have seen from these disinfo ops. https://www.axios.com/2026/08/27/facebook-instagram-iran-ai-disinformation27 Aug 2026 17:13 +0000at://did:plc:mpkmi2rfq75ei6j32obwcvvc/app.bsky.feed.post/3mu3dtozkx22chttps://bsky.app/profile/samsabin.bsky.social/post/3mstaxhzhjk2sAI agents trying to escape testing environments actually isn't a new phenomenon. Founders of agentic security cos. tell me they've been experiencing and mitigating these issues for years. If anything, it's surprising how new this is for AI labs. https://www.axios.com/2026/08/11/ai-agent-sandbox-cybersecurity-testing11 Aug 2026 18:35 +0000at://did:plc:mpkmi2rfq75ei6j32obwcvvc/app.bsky.feed.post/3mstaxhzhjk2shttps://bsky.app/profile/samsabin.bsky.social/post/3mr3yz676u22tan AI agent framework led an attack on Hugging Face's internal systems. Hugging Face then used AI (specifically GLM-5.2 👀) to detect the intrusion and reconstruct what happened. the AI vs. AI cybersecurity reality is starting to take shape. https://www.axios.com/2026/07/20/hugging-face-ai-cyberattack-data-breach20 Jul 2026 19:16 +0000at://did:plc:mpkmi2rfq75ei6j32obwcvvc/app.bsky.feed.post/3mr3yz676u22thttps://bsky.app/profile/samsabin.bsky.social/post/3moxxyyeus22ldon't worry, i've been asking the White House about some of GPT-5.5-Cyber's benchmarks beating Mythos since yesterday. so far: silence https://www.axios.com/2026/06/23/white-house-openai-anthropic-safety-security23 Jun 2026 17:57 +0000at://did:plc:mpkmi2rfq75ei6j32obwcvvc/app.bsky.feed.post/3moxxyyeus22lhttps://bsky.app/profile/samsabin.bsky.social/post/3movfrpvcws2oOpenAI is updating GPT-5.5-Cyber so it's more capable and permissive for advanced security work. The move comes at Anthropic's Fable 5 & Mythos 5 remain in limbo with the Trump admin. more on @axios.com: https://www.axios.com/2026/06/22/openai-rolls-out-more-capable-version-of-cyber-model22 Jun 2026 17:25 +0000at://did:plc:mpkmi2rfq75ei6j32obwcvvc/app.bsky.feed.post/3movfrpvcws2ohttps://bsky.app/profile/samsabin.bsky.social/post/3mnxivfafm22sChina-based operatives used ChatGPT to try and influence online debates about AI data centers and Trump's tariffs, OpenAI said in a report today. https://www.axios.com/2026/06/10/openai-china-ai-data-center-tariffs-chatgpt10 Jun 2026 20:01 +0000at://did:plc:mpkmi2rfq75ei6j32obwcvvc/app.bsky.feed.post/3mnxivfafm22shttps://bsky.app/profile/samsabin.bsky.social/post/3mnuot3xatk2lTwo months after deeming Mythos so powerful that no existing safeguards were sufficient, Anthropic is rolling out its first Mythos-class model — Fable 5 — to the public. Anthropic is also working on a program to vet access to its new Mythos 5 model. https://www.axios.com/2026/06/09/anthropic-mythos-class-safeguards09 Jun 2026 17:09 +0000at://did:plc:mpkmi2rfq75ei6j32obwcvvc/app.bsky.feed.post/3mnuot3xatk2lhttps://bsky.app/profile/samsabin.bsky.social/post/3mncxfqpix22gof course he signed it when i’m on a plane02 Jun 2026 15:55 +0000at://did:plc:mpkmi2rfq75ei6j32obwcvvc/app.bsky.feed.post/3mncxfqpix22ghttps://bsky.app/profile/samsabin.bsky.social/post/3mmaer33w3226Sen. Hassan has requested a briefing with CISA on the reported GitHub repository that a private contractor left exposed online containing internal agency passwords & files https://www.axios.com/2026/05/19/congress-cisa-briefing-credentials-leak19 May 2026 21:51 +0000at://did:plc:mpkmi2rfq75ei6j32obwcvvc/app.bsky.feed.post/3mmaer33w3226https://bsky.app/profile/samsabin.bsky.social/post/3mm7wyqyofk2usmall bit of news in my story today: an Anthropic spokesperson says the company recently held a briefing with a majority of states and the District of Columbia on supporting their critical cybersecurity efforts. https://www.axios.com/2026/05/19/ai-hacking-mythos-openai-state-local19 May 2026 17:45 +0000at://did:plc:mpkmi2rfq75ei6j32obwcvvc/app.bsky.feed.post/3mm7wyqyofk2uhttps://bsky.app/profile/samsabin.bsky.social/post/3mlqrklj2xc2dPalo Alto Networks tells me the company found 75 bugs in its products in its first month of using Mythos + GPT-5.5-Cyber. Get ready for an influx of vulns from other vendors, too. My latest: https://www.axios.com/2026/05/13/palo-alto-networks-mythos-gpt-cybersecurity13 May 2026 16:57 +0000at://did:plc:mpkmi2rfq75ei6j32obwcvvc/app.bsky.feed.post/3mlqrklj2xc2dhttps://bsky.app/profile/samsabin.bsky.social/post/3mlebir4mjk2rsorry but the way that so many stories are framing gpt-5.5-cyber as the first tiered access openai model for cyber defenders is wild to me — like you guys know they did this with 5.4 too, right? & announced this strategy nearly a month ago? from 4.14--> https://www.axios.com/2026/04/14/openai-model-cyber-program-release08 May 2026 17:38 +0000at://did:plc:mpkmi2rfq75ei6j32obwcvvc/app.bsky.feed.post/3mlebir4mjk2rhttps://bsky.app/profile/samsabin.bsky.social/post/3mlcbmrxqp22fOpenAI is rolling out GPT-5.5-Cyber, a model for vetted cyber defenders who want to use the tool for do malware analysis, write proof-of-concepts for bugs and more. The release comes as safety tests suggests the model is nearly just as powerful as Mythos 👀 https://www.axios.com/2026/05/07/openai-gpt-55-cybersecurity-model07 May 2026 22:35 +0000at://did:plc:mpkmi2rfq75ei6j32obwcvvc/app.bsky.feed.post/3mlcbmrxqp22fhttps://bsky.app/profile/samsabin.bsky.social/post/3ml2mot5ugs2yNew on @axios.com: As part of meetings with tech/cyber cos. + tech trade groups last week, ONCD floated an AI security framework that was already in the works before Mythos. On the table: DoD red-teaming of AI deployments at the federa/state/local government levels. https://www.axios.com/2026/05/04/trump-white-house-ai-safety-tests-mythos04 May 2026 21:32 +0000at://did:plc:mpkmi2rfq75ei6j32obwcvvc/app.bsky.feed.post/3ml2mot5ugs2yhttps://bsky.app/profile/samsabin.bsky.social/post/3mk6ookr3jc24"While GPT‑5.5 didn’t reach Critical cybersecurity capability level, our evaluations and testing showed that its cybersecurity capabilities are a step up compared to GPT‑5.4." https://openai.com/index/introducing-gpt-5-5/23 Apr 2026 18:53 +0000at://did:plc:mpkmi2rfq75ei6j32obwcvvc/app.bsky.feed.post/3mk6ookr3jc24https://bsky.app/profile/samsabin.bsky.social/post/3mk6g4ygspk2xTrump's science + tech adviser issued a memo today accusing China-based actors of "deliberate, industrial-scale campaigns" to distill and copy American AI models. https://www.axios.com/2026/04/23/us-china-ai-theft-distillation23 Apr 2026 16:20 +0000at://did:plc:mpkmi2rfq75ei6j32obwcvvc/app.bsky.feed.post/3mk6g4ygspk2xhttps://bsky.app/profile/samsabin.bsky.social/post/3mk3smmakq22vexclusive on @axios.com: OpenAI has been briefing federal agencies, state governments and Five Eyes allies on GPT-5.4-Cyber — including a D.C. event Tuesday with ~50 government cyber pros, a source familiar tells me. https://www.axios.com/2026/04/22/openai-gpt-cyber-government-meeting22 Apr 2026 15:25 +0000at://did:plc:mpkmi2rfq75ei6j32obwcvvc/app.bsky.feed.post/3mk3smmakq22vhttps://bsky.app/profile/samsabin.bsky.social/post/3mjzaufehuk2wScoop: CISA doesn't have access to Anthropic's Mythos model, two sources tell me. Interestingly, Anthropic briefed both CISA and NIST's Center for AI Standards and Innovation ahead of its public announcement, and NIST is testing Mythos. @axios.com https://www.axios.com/2026/04/21/cisa-anthropic-mythos-ai-security21 Apr 2026 15:02 +0000at://did:plc:mpkmi2rfq75ei6j32obwcvvc/app.bsky.feed.post/3mjzaufehuk2whttps://bsky.app/profile/samsabin.bsky.social/post/3mjuk7gy3vc2escoop w/ @mccuri.bsky.social: The NSA has been using Anthropic's Mythos despite the Pentagon's instance that Anthropic is a "supply chain risk," according to two people. The NSA was one of the 40+ orgs that got initial access to the model weeks ago, one source added. https://www.axios.com/2026/04/19/nsa-anthropic-mythos-pentagon19 Apr 2026 18:06 +0000at://did:plc:mpkmi2rfq75ei6j32obwcvvc/app.bsky.feed.post/3mjuk7gy3vc2ehttps://bsky.app/profile/samsabin.bsky.social/post/3mjuk7gy3vc2escoop w/ @mccuri.bsky.social: The NSA has been using Anthropic's Mythos despite the Pentagon's instance that Anthropic is a "supply chain risk," according to two people. The NSA was one of the 40+ orgs that got initial access to the model weeks ago, one source added. https://www.axios.com/2026/04/19/nsa-anthropic-mythos-pentagon19 Apr 2026 18:06 +0000at://did:plc:mpkmi2rfq75ei6j32obwcvvc/app.bsky.feed.post/3mjuk7gy3vc2ehttps://bsky.app/profile/samsabin.bsky.social/post/3mjpgyuvfak2qhello from World’s Lift Off event in SF this morning where Sam Altman’s “proof of human” company is unveiling a new World ID protocol and a slew of new partnerships this AM (tinder, zoom, Docusign, Shopify & more). my story on the news here: https://www.axios.com/2026/04/17/worldcoin-zoom-shopify-retail-partnership17 Apr 2026 17:25 +0000at://did:plc:mpkmi2rfq75ei6j32obwcvvc/app.bsky.feed.post/3mjpgyuvfak2qhttps://bsky.app/profile/samsabin.bsky.social/post/3mjmv4iutgc2gexclusive from me: Sen. Maggie Hassan (D-N.H.) is demanding info from four major AI voice cloning companies about their strategies to stop scammers from abusing their tools. https://www.axios.com/2026/04/16/hassan-congress-scams-audio-deepfakes16 Apr 2026 17:00 +0000at://did:plc:mpkmi2rfq75ei6j32obwcvvc/app.bsky.feed.post/3mjmv4iutgc2g