AIVIO News — The state of artificial intelligence, every day.

Trump administration backs OpenAI in copyright fight with The New York Times

Wed, 02 Sep 2026 22:47:30 GMT

The Trump administration filed a 20-page amicus brief backing OpenAI's fair use defense, arguing that limiting AI training would harm U.S. competitiveness in artificial intelligence.

The Trump administration filed a 20-page brief in The New York Times' copyright lawsuit against OpenAI, backing OpenAI's position.
The case is in the U.S. District Court for the Southern District of New York, where no ruling has been made yet.
Judge William Alsup ordered Anthropic to pay $1.5 billion last year for using pirated books, not for AI training itself.
The brief cites a Trump executive order calling for the U.S. to retain global leadership in artificial intelligence.

Anthropic drops Claude Fable 5.1 with cheaper pricing and lighter restrictions

Wed, 02 Sep 2026 22:47:30 GMT

Claude Fable 5.1 costs up to 45 percent less for complex agentic tasks and blocks fewer basic biology questions than its predecessor, Anthropic says.

Fable 5.1 costs about 25 percent less than Fable 5 typically, and up to 45 percent less for complex agentic tasks, Anthropic says.
The savings come mainly from cheaper pricing on cached data the model has already processed and stored.
Claude Fable 5.1 is available on all platforms now; Mythos 5.1 is limited to Project Glasswing participants.
Anthropic says Enterprise Frontier Safeguards, which store customer data on the customer's own servers, roll out later this fall.

OpenAI's Astra model raises alarm with harder-to-monitor reasoning

Wed, 02 Sep 2026 22:47:30 GMT

OpenAI's Astra uses a looping technique called recurrent depth that processes queries multiple times, and safety researchers warn wider use could make AI reasoning impossible to monitor.

OpenAI's Astra model uses a technique called recurrent depth, also known as opaque recurrence, The Information reported Tuesday.
The technique processes a query multiple times in a loop rather than in a single linear chain of thought.
Redwood Research CEO Buck Shlegeris and AI safety advocate Zvi Mowshowitz both raised alarms about the technique's impact on monitoring.
OpenAI chief scientist Jakub Pachocki said preserving legible chains of thought remains a core research goal.

Palo Alto Networks buys IT automation startup Console for $500 million

Wed, 02 Sep 2026 22:47:30 GMT

The $500 million cash-and-stock deal folds Console's AI agents into Palo Alto Networks' Cortex platform, marking the cybersecurity giant's seventh acquisition of 2026.

Console raised $29 million total, including a $6.2 million seed round led by Thrive Capital.
PitchBook valued Console at $157 million before Palo Alto Networks acquired it for $500 million.
Rival startup Serval reached a $1 billion valuation after a Sequoia-led round in December 2025.
This is Palo Alto Networks' seventh acquisition of 2026, according to PitchBook.

Google DeepMind launches Gemini 3.8 Flash and a cybersecurity version

Wed, 02 Sep 2026 22:27:49 GMT

The new model pushes deeper multi-step reasoning and tool use for coding agents, while the cyber-focused variant is limited to vetted defenders under a new access program called Fairwind.

Gemini 3.8 Flash costs $0.75 per million input tokens and $3.75 per million output tokens, matching 3.7 Flash's price.
Gemini 3.8 Flash scored 54.9% on the HLE-Verified reasoning benchmark, according to Google DeepMind.
Gemini 3.8 Flash Cyber found more than 70% of vulnerabilities in an internal test spanning 20 programming languages.
Chrome's security team says 3.8 Flash Cyber produced 2.6 times more correct patches than the best commercial models it tested.

OpenAI says Astra is first model to cross Critical cybersecurity threshold

Wed, 02 Sep 2026 22:27:49 GMT

The classification, made under OpenAI's internal Preparedness Framework, triggers stronger safeguards before release, though the company has not detailed what those safeguards involve.

Astra is the first OpenAI model to meet the Critical cybersecurity capability threshold.
The classification falls under OpenAI's Preparedness Framework.
OpenAI says Astra will have stronger safeguards for its release.