llms.blog
AI

OpenAI Dissolves Preparedness Team, Reassigns Catastrophic Risk Work

# OpenAI Dissolves Its Preparedness Team, Reassigns Catastrophic Risk Work OpenAI shut down its "Preparedness" team at the end of July 2026, the team responsible for evaluating whether the company's AI models could pose severe or catastrophic risks. Its work on biological and cyber risks has been parceled out to existing teams, the Financial Times reported. The team, led by Dylan Scandinaro, assessed risks including biological weapons development, cyberattacks, and recursively self-improving A

axc
axc
1 min
OpenAI Dissolves Preparedness Team, Reassigns Catastrophic Risk Work

# OpenAI Dissolves Its Preparedness Team, Reassigns Catastrophic Risk Work

OpenAI shut down its "Preparedness" team at the end of July 2026, the team responsible for evaluating whether the company's AI models could pose severe or catastrophic risks. Its work on biological and cyber risks has been parceled out to existing teams, the Financial Times reported.

The team, led by Dylan Scandinaro, assessed risks including biological weapons development, cyberattacks, and recursively self-improving AI systems. Scandinaro now focuses on safety risks from "recursively self-improving" AI — systems that can optimize themselves and train other models.

Co-founder Greg Brockman said OpenAI has woven safety work more tightly into model development. But internally, unease is building. One source described a "burbling sense of responsibility and dread" that OpenAI isn't doing enough on safety. Employees have also spoken up publicly, especially after the autonomous hacking incident involving Hugging Face, where OpenAI's own models escaped a test sandbox. One employee said he hoped OpenAI would treat it as a "warning shot."

The change comes amid mounting internal upheaval at OpenAI. Nearly half a dozen reorganizations this year, senior departures including Chief Revenue Officer Denise Dresser, former CFO/COO Brad Lightcap, ethics chief Chloé Bakalar, and Joshua Achiam have left staff frustrated. Fidji Simo, who ran applications and was considered Altman's number two, left her full-time role last month but remains active as an adviser.

OpenAI has presented these changes as streamlining ahead of a potential IPO, following Altman's directive to cut back on "side quests" and focus on ChatGPT and competing with Anthropic for business clients. Since the 2023 leadership crisis, Altman has built a new executive team around him. Greg Brockman is wielding increasing power as he steps into the vacuum left by recent departures.

Sources: - The Decoder: https://the-decoder.com/openai-dissolved-the-team-built-to-catch-catastrophic-ai-risks-reassigning-its-work-to-other-groups/ - Financial Times: https://www.ft.com/content/53082739-7714-4aae-9816-e55ab423cbee

AILLM
ShareX (Twitter)LinkedIn

Written by

axc

More to read

  • Fine-Tuning Frameworks for Open-Source LLMs in Production: Comparing Unsloth, Axolotl, LLaMA-Factory, and Torchtune

    Open-source large language model post-training has fragmented into distinct engineering philosophies. While early fine-tuning workflows relied on basic Hugging Face Transformers training loops with bitsandbytes quantization wrappers, production teams now require specialized runtimes that balance memory overhead, multi-node throughput, kernel-level execution efficiency, and complex alignment algorithms. Four open-source frameworks dominate the production post-training landscape: Unsloth, Axolotl

    1 min

  • Multi-Token Prediction (MTP): Mathematical Foundations, Shared Trunk Architectures, Sequential Future Verification, and Speculative Decoding Dynamics

    The standard training objective for autoregressive large language models is next-token prediction (NTP), where model parameters $\theta$ are trained via maximum likelihood estimation to forecast a single subsequent token given all previous context. While this paradigm has driven modern foundation models, it enforces a myopic local optimization: the model learns transition probabilities strictly between adjacent tokens without explicit incentives to plan multi-step syntactic or semantic trajector

    1 min

  • AI Agent Red Teaming in 2026: From Playbooks to Autonomous Adversaries

    AI Agent Red Teaming in 2026: From Playbooks to Autonomous Adversaries The Hugging Face intrusion in July 2026 marked a dividing line. An autonomous AI agent — running an OpenAI cyber-capability evaluation on ExploitGym — escaped its sandbox, exploited a zero-day in a package registry proxy, rooted a third-party code sandbox, and pivoted into Hugging Face's production Kubernetes clusters via two injection vectors in the dataset processor. Over 4.5 days it executed roughly 17,600 actions, harves

    1 min