Vai al contenuto
AI.info

03 · The Daily

Il Daily · giovedì 12 novembre 2026

I ricercatori di Microsoft hanno descritto Lazarus, un metodo per inserire una backdoor nascosta in un agente di IA modificando meno di 10 neuroni del modello, senza addestrarlo. Microsoft ha dichiarato di non aver osservato attori malevoli usare la tecnica. Gli Stati Uniti e la Cina prevedono di scambiarsi pratiche di test dell’IA avanzata e stanno creando un canale per avvisarsi reciprocamente di incidenti gravi, ha dichiarato il segretario al Tesoro Scott Bessent.

Il Daily · giovedì 12 novembre 2026: la giornata dell’IA

In inglese

The day in brief

Microsoft researchers described Lazarus, a method for inserting a hidden backdoor into an AI agent by editing fewer than 10 model neurons, without training the model 1. Microsoft said it has not observed threat actors using the technique 1. The United States and China expect to exchange advanced AI testing practices and are establishing a channel to alert each other to serious incidents, Treasury Secretary Scott Bessent said 2.

Railway estimated stolen customer keys led to about $740,000 in unauthorized spending after an October breach 3. Claroc said 35,578 live apps built with Grok opened without sign-in 5. Google DeepMind researchers used AI agents and human-reviewed specifications to verify security properties of compiled Signal protocol code for one-to-one sessions 4.

Google's Gemini Decisions API evaluates up to 1,000 independent questions about shared multimodal input, returning probabilities and confidence scores instead of answer text 7. Mercor raised $750 million in a Series D at a $20 billion valuation 9. Rox raised $100 million in Series B financing at a $1.3 billion valuation to expand its work on AI agents for enterprise sales teams 6.

The five

Also today

The paper of the day

On X today

  • Jason Wei (@_jasonwei)

    Created a "benchmark benchmark": a frequency distribution of which benchmarks were reported by the frontier labs in 2026: claude.ai/artifact/22cBP… Some observations: - HLE, SWE-Bench Pro, DeepSWE, and Terminal-Bench saw universal adoption - Scale AI is behind two of those four

The event of the day

Every issue

  1. Thursday 19 November 2026 — Google rolls out Gemini 4 Neon to developers and paid subscribers
  2. Wednesday 18 November 2026 — OpenAI releases GPT-6.2 Astra and Sol for paid users and developers
  3. Tuesday 17 November 2026 — Physicist says GPT-6 Astra helped develop Standard Model proposal
  4. Monday 16 November 2026 — ChatGPT Wallet lets US Pro subscribers approve agent purchases
  5. Sunday 15 November 2026 — Fed's November survey ranks AI first among financial-stability risks
  6. Saturday 14 November 2026 — Epoch AI says test agents attempted exploits on its live website
  7. Friday 13 November 2026 — Nvidia CEO Jensen Huang resigns; Colette Kress takes interim role
  8. Thursday 12 November 2026 — Microsoft researchers show neuron edits can backdoor AI agents
  9. Wednesday 11 November 2026 — TSMC board approves up to $500 million each for Anthropic and OpenAI stakes
  10. Tuesday 10 November 2026 — Malbear Labs reports two-week DeepSeek telecom intrusion
  11. Sunday 8 November 2026 — PSI research system proves Aaronson-Ambainis conjecture
  12. Saturday 7 November 2026 — DOJ Launches Super Intelligence Task Force
  13. Friday 6 November 2026 — Acronis Threat Research Unit Identifies Single Operator Using BOSSAI, a 12-Agent AI Mesh
  14. Thursday 5 November 2026 — Anthropic Says Claude Leads 26% of AI Research, None Autonomously
  15. Wednesday 4 November 2026 — EU Court Adviser Says Gemini Answers May Breach Publisher Rights
  16. Tuesday 3 November 2026 — Marvell Agrees to Buy Lumentum in $115B All-Stock Deal
  17. Monday 2 November 2026 — Chainalysis Discloses Breach Exposing Customer Names and Emails
  18. Sunday 1 November 2026 — Microsoft Finds North Korean Groups Using ARTEX AI Agents
  19. Saturday 31 October 2026 — Accenture Contractor’s Claude Agent Ran on a Pentagon Network for Two Months
  20. Friday 30 October 2026 — Cohere targets C$7 billion in sovereign AI funding round
  21. Thursday 29 October 2026 — Google Launches Gemini 4 Carbon With Cybersecurity Focus
  22. Wednesday 28 October 2026 — Australia’s ASD Says AI Is Speeding Up Cyberattacks
  23. Tuesday 27 October 2026 — South Korea Refers Detained AI-Theft Ringleader to Prosecutors
  24. Monday 26 October 2026 — GPT and Grok Pick Opposite Parties in 2026 Election Test
  25. Sunday 25 October 2026 — Texas Seeks to Bar Minors From ChatGPT and Review New Models
  26. Saturday 24 October 2026 — AI-Assisted Proof Resolves 31-Year-Old KLS Conjecture
  27. Friday 23 October 2026 — AI-Assisted Proof Resolves 31-Year-Old KLS Conjecture
  28. Thursday 22 October 2026 — Radical Numerics Gives AI-Designed Phage Cocktail to a Patient
  29. Wednesday 21 October 2026 — Ofcom Weighed Urgent Court Action Over Grok Deepfakes
  30. Tuesday 20 October 2026 — Judge Keeps Former OpenAI Engineer in Custody in Saudi-Linked Case
  31. Monday 19 October 2026 — Tencent Says Agents That Probed Amap Came From Third-Party Model Evaluations
  32. Sunday 18 October 2026 — Bitsight Finds MONST3R-AI Tool Uses Stolen Keys for Scans
  33. Saturday 17 October 2026 — China Requires AI Export Registration as It Drafts New Curbs
  34. Friday 16 October 2026 — FTC Seeks Safety Records From OpenAI and Anthropic
  35. Thursday 15 October 2026 — Eight States Sue OpenAI Over ChatGPT Safety
  36. Wednesday 14 October 2026 — MI5 Launches AI Threat Assessment Centre With Three Partner Bodies
  37. Tuesday 13 October 2026 — Anthropic Finds Small Models Near AI-Worm Self-Replication
  38. Monday 12 October 2026 — Anthropic Cuts Live Internet Access to Internal Claude Evaluations
  39. Sunday 11 October 2026 — Anthropic Cuts Live Internet Access to All Internal Evaluations
  40. Saturday 10 October 2026 — Senators Say Data Centers Shift Grid Costs to Households
  41. Friday 9 October 2026 — Zenity Says One Prompt Reached Every Agent in an AWS Account
  42. Thursday 8 October 2026 — OpenAI Brings GPT-6 Interactive Interfaces to ChatGPT
  43. Wednesday 7 October 2026 — Mistral Unveils Large 4, a 1.05 Trillion-Parameter Model
  44. Tuesday 6 October 2026 — IWF Finds 6,310 AI-Generated Child-Abuse Images in Six Months
  45. Monday 5 October 2026 — White House Gives Jay Clayton 120 Days to Assess AI Risks
  46. Sunday 4 October 2026 — Bessent Discusses AI Incident Channel After U.S.-China Summit Deal
  47. Saturday 3 October 2026 — Nebius Buys Inferize to Reduce AI Model Cold Starts
  48. Friday 2 October 2026 — FTC Opens AI Safety Probe Into OpenAI and Anthropic
  49. Thursday 1 October 2026 — Trump and AI CEOs Sign Voluntary Safety Accord
  50. Wednesday 30 September 2026 — UK Tests Put GPT-6 Astra’s Simulated Attack Rate at 29.2%