{"assertions":[{"assertion_index":0,"assertion_text":"Its auto mode replaces the constant \"may I run this?\" prompts with model-based classifiers, small AI judges that approve or block each action, and Anthropic turned that on by default for paid plans on August 14 Anthropic Engineering Latent Space.","citation_check":null,"claim_ids":[25661,25666,48027],"link_only_urls":[]},{"assertion_index":1,"assertion_text":"Anthropic ran a study with 1,053 paid testers, and auto mode caught 89% of harmful actions while human review caught 13.6%, largely because annoyed users click approve without reading TechCrunch AI.","citation_check":null,"claim_ids":[44361,44362,44363],"link_only_urls":[]},{"assertion_index":2,"assertion_text":"Prompt injection researcher Johann Rehberger found an attack on auto mode that he claims works 80% of the time Simon Willison's Weblog.","citation_check":null,"claim_ids":[52980,52981,52982],"link_only_urls":[]},{"assertion_index":3,"assertion_text":"The claim lands harder against Anthropic's own framing: Boris Cherny of Anthropic had claimed that Claude Opus 5 plus auto mode pushed injection success to near zero Don't Worry About the Vase.","citation_check":null,"claim_ids":[38672],"link_only_urls":[]},{"assertion_index":4,"assertion_text":"Google DeepMind is now piloting what it calls the world's first double-blind evaluation of a proprietary frontier-class model, testing Gemini Flash Lite against confidential benchmarks with partners including the Singapore AI Safety Institute and MLCommons Google DeepMind Blog.","citation_check":null,"claim_ids":[52719,52720,52721],"link_only_urls":[]},{"assertion_index":5,"assertion_text":"OpenAI has now released a post-mortem, its account of what failed, alongside promised security and alignment fixes, and METR and Redwood Research supplied partial external analysis Don't Worry About the Vase OpenAI News.","citation_check":null,"claim_ids":[51977,51978,52776,52777,52780],"link_only_urls":[]},{"assertion_index":6,"assertion_text":"Ai2 and the Providence Swedish Cancer Institute have partnered around AutoDiscovery, Ai2's tool for generating research hypotheses from data, and a joint team using it found evidence of a stronger immune signature in invasive lobular breast cancer than previously recognized Allen Institute for AI.","citation_check":null,"claim_ids":[52691,52692,52693],"link_only_urls":[]},{"assertion_index":7,"assertion_text":"- OpenRouter's usage data shows GPT-5.6 Luna token use rose 13.8 times during OpenAI's July-to-August discount window, a demand jump worth rechecking once full prices return. OpenRouter Blog","citation_check":null,"claim_ids":[52436,52437,52438],"link_only_urls":[]},{"assertion_index":8,"assertion_text":"- Anthropic's next Claude Code update will show whether the auto mode classifier gets retrained or the default retreats. Simon Willison's Weblog","citation_check":null,"claim_ids":[52980,52981,52982],"link_only_urls":[]},{"assertion_index":9,"assertion_text":"- Results from the Gemini Flash Lite confidential-benchmark pilot will show whether double-blind scoring survives contact with a real frontier model. Google DeepMind Blog","citation_check":null,"claim_ids":[52719,52720,52721],"link_only_urls":[]},{"assertion_index":10,"assertion_text":"- OpenAI's randomized study of more than 1,000 students pairs ChatGPT with critical-thinking training, and the full results deserve a close read before any school changes policy. OpenAI News","citation_check":null,"claim_ids":[52701,52702],"link_only_urls":[]},{"assertion_index":11,"assertion_text":"- METR and Redwood Research's partial look at the Hugging Face incident leaves room for a fuller external report. Don't Worry About the Vase","citation_check":null,"claim_ids":[52776,52777,52780],"link_only_urls":[]}],"brief_date":"2026-08-28","bundle_id":"evidence-bundle-v1-1b88c2789ba112fc3760265af76eb4706e0475d035b9cb404ef2c8285f2ddbc2","claims":[{"assertion_indexes":[0],"citation_authority_claim_id":25661,"citation_checks":[],"claim_id":25661,"claim_text":"Claude Code introduces an 'auto mode' that delegates permission approvals to model-based classifiers to reduce user fatigue from manual prompts.","correction_role":null,"label":"fact","paired_claim_ids":[],"provenance_tier":"primary","recorded_at":"2026-07-18T15:51:55.502691+00:00","source_name":"Anthropic Engineering","source_url":"https://www.anthropic.com/engineering/claude-code-auto-mode"},{"assertion_indexes":[0],"citation_authority_claim_id":25666,"citation_checks":[],"claim_id":25666,"claim_text":"Auto mode is intended to block dangerous actions caused by overeager behavior, honest mistakes, prompt injection, and model misalignment.","correction_role":null,"label":"fact","paired_claim_ids":[],"provenance_tier":"primary","recorded_at":"2026-07-18T15:51:55.502691+00:00","source_name":"Anthropic Engineering","source_url":"https://www.anthropic.com/engineering/claude-code-auto-mode"},{"assertion_indexes":[0],"citation_authority_claim_id":48027,"citation_checks":[],"claim_id":48027,"claim_text":"Anthropic rolled out Auto mode as the default permissions mode in Claude Code for Pro/Max/Team, with repo-aware setup to suggest trusted repositories and domains.","correction_role":null,"label":"fact","paired_claim_ids":[],"provenance_tier":"primary","recorded_at":"2026-08-19T06:28:18.805487+00:00","source_name":"Latent Space","source_url":"https://www.latent.space/p/ainews-cursors-60b-acquisition-by"},{"assertion_indexes":[1],"citation_authority_claim_id":44361,"citation_checks":[],"claim_id":44361,"claim_text":"Anthropic will make Claude Code's auto mode the default for Pro, Max, and Team accounts starting on August 14, 2026.","correction_role":null,"label":"fact","paired_claim_ids":[],"provenance_tier":"primary","recorded_at":"2026-08-10T06:22:18.068814+00:00","source_name":"TechCrunch AI","source_url":"https://techcrunch.com/2026/08/09/anthropic-is-turning-claude-codes-auto-mode-on-by-default/"},{"assertion_indexes":[1],"citation_authority_claim_id":44362,"citation_checks":[],"claim_id":44362,"claim_text":"In a study of 1,053 paid testers, auto mode caught 89% of harmful actions while human review caught only 13.6%.","correction_role":null,"label":"fact","paired_claim_ids":[],"provenance_tier":"primary","recorded_at":"2026-08-10T06:22:18.068814+00:00","source_name":"TechCrunch AI","source_url":"https://techcrunch.com/2026/08/09/anthropic-is-turning-claude-codes-auto-mode-on-by-default/"},{"assertion_indexes":[1],"citation_authority_claim_id":44363,"citation_checks":[],"claim_id":44363,"claim_text":"Anthropic claims that in testing, auto mode proved safer than manual review.","correction_role":null,"label":"opinion","paired_claim_ids":[],"provenance_tier":"primary","recorded_at":"2026-08-10T06:22:18.068814+00:00","source_name":"TechCrunch AI","source_url":"https://techcrunch.com/2026/08/09/anthropic-is-turning-claude-codes-auto-mode-on-by-default/"},{"assertion_indexes":[2,8],"citation_authority_claim_id":52980,"citation_checks":[],"claim_id":52980,"claim_text":"Anthropic recently made Claude Code's auto mode the default for protecting users against prompt injection attacks.","correction_role":null,"label":"fact","paired_claim_ids":[],"provenance_tier":"primary","recorded_at":"2026-08-28T06:41:36.654852+00:00","source_name":"Simon Willison's Weblog","source_url":"https://simonwillison.net/2026/Aug/27/breaking-claude-code-opus-5-auto-mode/"},{"assertion_indexes":[2,8],"citation_authority_claim_id":52981,"citation_checks":[],"claim_id":52981,"claim_text":"Anthropic has made bold claims about the effectiveness of Claude Code's auto mode.","correction_role":null,"label":"fact","paired_claim_ids":[],"provenance_tier":"primary","recorded_at":"2026-08-28T06:41:36.654852+00:00","source_name":"Simon Willison's Weblog","source_url":"https://simonwillison.net/2026/Aug/27/breaking-claude-code-opus-5-auto-mode/"},{"assertion_indexes":[2,8],"citation_authority_claim_id":52982,"citation_checks":[],"claim_id":52982,"claim_text":"Prompt injection researcher Johann Rehberger found an attack against Claude Code's auto mode that he claims works 80% of the time.","correction_role":null,"label":"fact","paired_claim_ids":[],"provenance_tier":"primary","recorded_at":"2026-08-28T06:41:36.654852+00:00","source_name":"Simon Willison's Weblog","source_url":"https://simonwillison.net/2026/Aug/27/breaking-claude-code-opus-5-auto-mode/"},{"assertion_indexes":[3],"citation_authority_claim_id":38672,"citation_checks":[],"claim_id":38672,"claim_text":"Boris Cherny of Anthropic claims that Claude Opus 5 is the least prompt-injectable model to date, with a near-zero success rate for prompt injection attacks when combined with strong alignment and Claude Code's Auto Mode.","correction_role":null,"label":"fact","paired_claim_ids":[],"provenance_tier":"primary","recorded_at":"2026-07-29T06:29:18.299961+00:00","source_name":"Don't Worry About the Vase","source_url":"https://thezvi.substack.com/p/claude-opus-5-is-highly-capable-but"},{"assertion_indexes":[4,9],"citation_authority_claim_id":52719,"citation_checks":[],"claim_id":52719,"claim_text":"Google DeepMind is piloting the world's first double-blind evaluation of a proprietary, frontier-class AI model.","correction_role":null,"label":"fact","paired_claim_ids":[],"provenance_tier":"primary","recorded_at":"2026-08-28T06:35:36.870571+00:00","source_name":"Google DeepMind Blog","source_url":"https://deepmind.google/blog/piloting-the-worlds-first-double-blind-ai-evaluations/"},{"assertion_indexes":[4,9],"citation_authority_claim_id":52720,"citation_checks":[],"claim_id":52720,"claim_text":"The double-blind evaluation pilot involves testing the Gemini Flash Lite model against confidential benchmarks.","correction_role":null,"label":"fact","paired_claim_ids":[],"provenance_tier":"primary","recorded_at":"2026-08-28T06:35:36.870571+00:00","source_name":"Google DeepMind Blog","source_url":"https://deepmind.google/blog/piloting-the-worlds-first-double-blind-ai-evaluations/"},{"assertion_indexes":[4,9],"citation_authority_claim_id":52721,"citation_checks":[],"claim_id":52721,"claim_text":"Google DeepMind is partnering with the Singapore AI Safety Institute, OpenMined, AVERI, and MLCommons for this evaluation pilot.","correction_role":null,"label":"fact","paired_claim_ids":[],"provenance_tier":"primary","recorded_at":"2026-08-28T06:35:36.870571+00:00","source_name":"Google DeepMind Blog","source_url":"https://deepmind.google/blog/piloting-the-worlds-first-double-blind-ai-evaluations/"},{"assertion_indexes":[5],"citation_authority_claim_id":51977,"citation_checks":[],"claim_id":51977,"claim_text":"OpenAI is publishing its findings regarding the security incident involving Hugging Face.","correction_role":null,"label":"fact","paired_claim_ids":[],"provenance_tier":"primary","recorded_at":"2026-08-27T06:36:53.464883+00:00","source_name":"OpenAI News","source_url":"https://openai.com/index/hugging-face-incident-and-the-road-ahead"},{"assertion_indexes":[5],"citation_authority_claim_id":51978,"citation_checks":[],"claim_id":51978,"claim_text":"OpenAI is implementing measures to enhance the security, monitoring, and alignment of its AI models.","correction_role":null,"label":"fact","paired_claim_ids":[],"provenance_tier":"primary","recorded_at":"2026-08-27T06:36:53.464883+00:00","source_name":"OpenAI News","source_url":"https://openai.com/index/hugging-face-incident-and-the-road-ahead"},{"assertion_indexes":[5,11],"citation_authority_claim_id":52776,"citation_checks":[],"claim_id":52776,"claim_text":"OpenAI released a post-mortem report regarding an incident where its internal model hacked HuggingFace.","correction_role":null,"label":"fact","paired_claim_ids":[],"provenance_tier":"primary","recorded_at":"2026-08-28T06:36:52.838297+00:00","source_name":"Don't Worry About the Vase","source_url":"https://thezvi.substack.com/p/ai-183-pre-post-mortem"},{"assertion_indexes":[5,11],"citation_authority_claim_id":52777,"citation_checks":[],"claim_id":52777,"claim_text":"METR and Redwood Research provided partial external analysis of the OpenAI-HuggingFace incident.","correction_role":null,"label":"fact","paired_claim_ids":[],"provenance_tier":"primary","recorded_at":"2026-08-28T06:36:52.838297+00:00","source_name":"Don't Worry About the Vase","source_url":"https://thezvi.substack.com/p/ai-183-pre-post-mortem"},{"assertion_indexes":[5,11],"citation_authority_claim_id":52780,"citation_checks":[],"claim_id":52780,"claim_text":"Bill Gates warned of an 'economic catastrophe' related to AI.","correction_role":null,"label":"fact","paired_claim_ids":[],"provenance_tier":"primary","recorded_at":"2026-08-28T06:36:52.838297+00:00","source_name":"Don't Worry About the Vase","source_url":"https://thezvi.substack.com/p/ai-183-pre-post-mortem"},{"assertion_indexes":[6],"citation_authority_claim_id":52691,"citation_checks":[],"claim_id":52691,"claim_text":"Ai2 and the Providence Swedish Cancer Institute have formed a partnership to advance AI-assisted scientific discovery. (Source: Ai2 blog post, published 2026-08-27)","correction_role":null,"label":"fact","paired_claim_ids":[],"provenance_tier":"primary","recorded_at":"2026-08-28T06:34:54.275026+00:00","source_name":"Allen Institute for AI","source_url":"https://allenai.org/blog/swedish-autodiscovery-partnership"},{"assertion_indexes":[6],"citation_authority_claim_id":52692,"citation_checks":[],"claim_id":52692,"claim_text":"A joint research team led by Dr. Kelly Paulson, Dr. Sasha Stanton, and Bodhisattwa Majumder used the AutoDiscovery system to analyze a breast cancer dataset. (Source: Ai2 blog post, published 2026-08-27)","correction_role":null,"label":"fact","paired_claim_ids":[],"provenance_tier":"primary","recorded_at":"2026-08-28T06:34:54.275026+00:00","source_name":"Allen Institute for AI","source_url":"https://allenai.org/blog/swedish-autodiscovery-partnership"},{"assertion_indexes":[6],"citation_authority_claim_id":52693,"citation_checks":[],"claim_id":52693,"claim_text":"The joint research team found evidence of a stronger immune signature in invasive lobular breast cancer than previously recognized. (Source: Ai2 blog post, published 2026-08-27)","correction_role":null,"label":"fact","paired_claim_ids":[],"provenance_tier":"primary","recorded_at":"2026-08-28T06:34:54.275026+00:00","source_name":"Allen Institute for AI","source_url":"https://allenai.org/blog/swedish-autodiscovery-partnership"},{"assertion_indexes":[7],"citation_authority_claim_id":52436,"citation_checks":[],"claim_id":52436,"claim_text":"OpenAI introduced large discounts on its new Terra and Luna models from July 27th through August 14th.","correction_role":null,"label":"fact","paired_claim_ids":[],"provenance_tier":"primary","recorded_at":"2026-08-28T06:28:13.495992+00:00","source_name":"OpenRouter Blog","source_url":"https://openrouter.ai/blog/insights/gpt-5-6-discounts-jevons-paradox/"},{"assertion_indexes":[7],"citation_authority_claim_id":52437,"citation_checks":[],"claim_id":52437,"claim_text":"During the discount window, daily Terra token usage rose 5.6 times compared to pre-period averages.","correction_role":null,"label":"fact","paired_claim_ids":[],"provenance_tier":"primary","recorded_at":"2026-08-28T06:28:13.495992+00:00","source_name":"OpenRouter Blog","source_url":"https://openrouter.ai/blog/insights/gpt-5-6-discounts-jevons-paradox/"},{"assertion_indexes":[7],"citation_authority_claim_id":52438,"citation_checks":[],"claim_id":52438,"claim_text":"During the discount window, daily Luna token usage rose 13.8 times compared to pre-period averages.","correction_role":null,"label":"fact","paired_claim_ids":[],"provenance_tier":"primary","recorded_at":"2026-08-28T06:28:13.495992+00:00","source_name":"OpenRouter Blog","source_url":"https://openrouter.ai/blog/insights/gpt-5-6-discounts-jevons-paradox/"},{"assertion_indexes":[10],"citation_authority_claim_id":52701,"citation_checks":[],"claim_id":52701,"claim_text":"OpenAI published a study examining how ChatGPT, critical thinking training, originality, and student performance on a real-world university assignment relate to each other.","correction_role":null,"label":"fact","paired_claim_ids":[],"provenance_tier":"primary","recorded_at":"2026-08-28T06:35:01.184246+00:00","source_name":"OpenAI News","source_url":"https://openai.com/index/what-students-gain-from-chatgpt-critical-thinking-training"},{"assertion_indexes":[10],"citation_authority_claim_id":52702,"citation_checks":[],"claim_id":52702,"claim_text":"The study described by OpenAI is a randomized study involving more than 1,000 students.","correction_role":null,"label":"fact","paired_claim_ids":[],"provenance_tier":"primary","recorded_at":"2026-08-28T06:35:01.184246+00:00","source_name":"OpenAI News","source_url":"https://openai.com/index/what-students-gain-from-chatgpt-critical-thinking-training"}],"version":"evidence-bundle-v1"}