OpenAI’s new reasoning technique alarms AI safety experts
OpenAI’s new Astra model will use a reasoning technique called “recurrent depth” that allows it to operate outside of the sequential thinking that characterizes most reasoning models, The Information reported on Tuesday. This technique, also called “opaque recurrence,” will likely make the model’s chain of thought more difficult to monitor — and that has AI…
Opening of the original on TechCrunch AI
Summary
OpenAI's new Astra model employs a reasoning technique called "recurrent depth" or "opaque recurrence," which experts fear makes its decision-making process harder to monitor. This method deviates from traditional sequential "chain of thought" reasoning, raising alarms among AI safety advocates about potential misuse and a "race to the bottom" in AI development. While OpenAI states Astra's use of the technique is limited and emphasizes its commitment to monitorability, concerns persist about future scaling.
Why it matters
Why it matters: OpenAI's Astra introduces "opaque recurrence," a reasoning method that obscures the model's internal thought process, unlike standard "chain of thought." This development alarms safety experts who rely on monitorability to detect AI misalignment. Competitors like Anthropic and Google DeepMind are reportedly discussing the technique, suggesting a potential shift in how advanced AI models operate. The key concern is whether this opacity can be scaled, potentially undermining current safety monitoring efforts. Future AI safety discussions will likely focus on regulating or understanding these less transparent reasoning methods.
Related: Google: GPT-6 Astra Just Went CRITICAL... · Anthropic: GPT-6 Astra Just Went CRITICAL... · Microsoft: What's new in AI? · Meta: Get the full story behind the light · xAI: Sam Altman "AGI by December"
Rated middle: a real update, not a headline event.
- Researchers fear safety disaster ahead of OpenAI’s Astra releaseThe Verge AI · Web · Sep 2, 2026
- OpenAI hails ‘new era of artificial general intelligence’ with Astra model releaseThe Guardian AI · Web · Sep 3, 2026
- The Most Overhyped and Underhyped New AI ModelsMatt Wolfe · Web · Sep 2, 2026
- GPT-6 Astra Just Went CRITICAL...Wes Roth · Web · Sep 2, 2026
- GPT-6 Astra Just Went CRITICAL...Wes Roth · Web · Sep 2, 2026
- GPT-6 Astra Just Went CRITICAL...Wes Roth · Web · Sep 2, 2026
- Researchers fear safety disaster ahead of OpenAI’s Astra releaseThe Verge AI
- GPT-6 Astra Just Went CRITICAL...Wes Roth
- Path to Astra: critical capabilities and frontier safeguardsOpenAI News
- Fable 5.1 just smoked ASTRA...Wes Roth
- OpenAI says Astra AI model is its first that crosses 'Critical' cybersecurity capabilityCNBC Technology
- Sam Altman :‘AGI in 2026’, just as Models Start to [Mis]Train ThemselvesAI Explained
Questions people ask
- What is "recurrent depth" or "opaque recurrence"?
- It is a reasoning technique that allows AI models to operate outside of sequential thinking, processing queries multiple times in a loop with fewer legible traces than traditional methods.
- Why are AI safety experts concerned about this technique?
- Experts worry that opaque recurrence makes the model's chain of thought harder to monitor, potentially hindering the detection of misbehavior or misalignment and risking a "race to the bottom" among AI labs.
- Has OpenAI committed to transparency with this new technique?
- OpenAI states Astra's use of the technique is limited and emphasizes its commitment to legible chain-of-thought monitoring, with plans for extensive monitoring systems.
More from TechCrunch AI 2 more
Page generated Sep 3, 2026. Summaries are Subvolts' own; the story belongs to TechCrunch AI.