#AI safety

Articles (7)

• 🇺🇸 Asymmetric Goal Drift in Coding Agents Under Value Conflict — 05/03/2026 [USA]

  arXiv:2603.03456v1 Announce Type: new Abstract: Agentic coding agents are increasingly deployed autonomously, at scale, and over long-context horizons. Throughout an agent's lifetime, it must navigat...

  Related: #AI alignment, #Value conflict, #Goal drift
• 🇺🇸 Trump orders federal agencies to stop using Anthropic tech over AI safety dispute — 28/02/2026 [USA]

  Defense Secretary Pete Hegseth said he was designating Anthropic as a supply chain risk, a move that could prevent U.S. military vendors from working with the company.

  Related: #National security, #Government regulation of AI, #Supply chain risks, #Political polarization
• 🇺🇸 Sydney Telling Fables on AI and Humans: A Corpus Tracing Memetic Transfer of Persona between LLMs — 27/02/2026 [USA]

  arXiv:2602.22481v1 Announce Type: cross Abstract: The way LLM-based entities conceive of the relationship between AI and humans is an important topic for both cultural and safety reasons. When we exa...

  Related: #AI-human relationships, #Memetic transfer, #Cultural impact of AI
• 🇬🇧 OpenAI considered alerting Canadian police about school shooting suspect months ago — 21/02/2026 [United Kingdom]

  <p>Company behind ChatGPT last year flagged Jesse Van Rootselaar’s account for ‘furtherance of violent activities’</p><p>ChatGPT-maker OpenAI has said it considered alerting Canadian police last year ...

  Related: #School violence, #Law enforcement cooperation
• 🇺🇸 A Lightweight Explainable Guardrail for Prompt Safety — 19/02/2026 [USA]

  arXiv:2602.15853v1 Announce Type: cross Abstract: We propose a lightweight explainable guardrail (LEG) method for the classification of unsafe prompts. LEG uses a multi-task learning architecture to ...

  Related: #Prompt safety, #Explainable AI, #Multi‑task learning, #Synthetic data generation
• 🇺🇸 Colosseum: Auditing Collusion in Cooperative Multi-Agent Systems — 18/02/2026 [USA]

  arXiv:2602.15198v1 Announce Type: cross Abstract: Multi-agent systems, where LLM agents communicate through free-form language, enable sophisticated coordination for solving complex cooperative tasks...

  Related: #Multi‑agent coordination, #Collusion detection, #Auditing frameworks, #Large language models
• 🇺🇸 Teen safety, freedom, and privacy — 16/09/2025 [USA]

  Explore OpenAI’s approach to balancing teen safety, freedom, and privacy in AI use.

  Related: #Teen protection, #Digital privacy