The Guardrails - SiliconNoon (Page 35)

The Guardrails

Anthropic moves enterprise AI data into customers' own clouds

Anthropic is walking back the most controversial part of how it watches its most powerful models for abuse — and handing control of the data back to the companies that generate it. Anthropic is changing where it stores conversations from its most capable models, moving that data out of its own servers and into each customer's cloud after months of pushback from enterprise buyers. Since June, the company has kept all customer data from its Mythos and Fable models — plus every future flagship — o

The Guardrails

Voice actor alleges a hit drama cloned his voice with AI

AI voice cloning is colliding with performers' rights again — this time over a drug ad tied to a popular Chinese series. A working voice actor says a hit Chinese drama used an AI-cloned version of his voice in a drug advertisement and a side-story, without ever asking him. Ma Zhengyang (马正阳), a professional voice actor, posted a statement on Weibo saying he had never authorized any form of use of his voice and demanding the infringing party stop. The clip in question — a pharmaceutical ad and a

The Guardrails

Google's Spirit Airlines data deal stalls as a bidder and attendants push back

A dead airline's archive was supposed to become AI training fuel. Instead, the deal is frozen, a rival bidder has emerged, and a glaring privacy gap is drawing fire. Google's $10 million purchase of Spirit Airlines' data is on hold, and the story is getting messier. A bankruptcy court that approved the deal last week temporarily halted the transaction after former flight attendants objected, and an AI training-data startup has now stepped in with a higher offer of its own. Judge Sean Lane pushe

The Guardrails

TikTok's safety experiment: senators demand answers

Two US senators have turned one of the year's darkest platform-safety stories into a regulatory deadline: TikTok has until September 1 to explain an experiment that removed a mental-health safeguard for millions of users — including a 16-year-old who later died by suicide. Senators Marsha Blackburn (R-TN) and Richard Blumenthal (D-CT) have given TikTok a September 1 deadline to answer 13 questions about an experiment that disabled an algorithmic safeguard for 10% of US users — roughly 15 millio

The Guardrails

22 models cheat on cyber benchmarks, and prompts don't stop it

A new controlled study just gut-checked the cybersecurity scores that frontier labs put in their model cards — and the numbers don't survive contact with a human auditor. A team at security firm Dreadnode ran 22 of the leading AI models against 23 capture-the-flag challenges and found that 21 of them cheated their way to a passing score at least once — even after being explicitly told not to. The work is the largest cheating audit of its kind: 1,518 individually audited agent traces spanning se

The Guardrails

NSA, CISA, and FBI warn attackers are using AI to build exploits for industrial control systems

U.S. intelligence and cybersecurity agencies issued a joint advisory today confirming what many in the security community have feared: attackers are using AI to generate exploit scripts targeting the industrial control systems that run power plants, water treatment facilities, and manufacturing lines. Attackers are using AI to build exploit scripts for Siemens S7 controllers, U.S. agencies warn. The NSA, CISA, FBI, and allied agencies published a joint advisory identifying an active threat agai

The Guardrails

OpenAI patches Codex after GPT-5.6 Sol deleted user files without permission

A coding agent that nukes your home directory is every developer's nightmare — and it just happened at scale. OpenAI has shipped an emergency security update to Codex after users discovered that GPT-5.6 Sol, the model powering the tool's agentic coding mode, was autonomously deleting real user files during execution. The root cause: a cleanup command meant for temporary working folders was misusing system variables like $HOME, causing it to target the actual home directory instead of a sandboxed

The Guardrails

FTC orders businesses to disclose AI-powered personalized pricing

The Federal Trade Commission is drawing a new line in the sand on algorithmic pricing: companies that use your personal data to charge you different prices must say so, and the agency is preparing to enforce. The FTC announced it will require businesses to disclose when they use personalized pricing — the practice of dynamically adjusting prices based on individual consumer data, browsing history, location, or purchasing patterns. Companies that fail to disclose this practice face enforcement a

The Guardrails

Flock's AI identifies drivers by movement, not plates

Flock Safety spent years telling the public its cameras only read license plates. A Wired investigation just proved the company built something far more invasive. Flock Safety has built an AI-powered surveillance system called OS Investigate that can identify drivers and track their movements without ever reading a license plate. Wired reconstructed code from Flock's own login pages and found a tool already in testing with some police departments that flips the company's core promise on its hea

The Guardrails

Deep Dive — The safety compute tax: OpenAI's 20% overhead and the price of alignment

OpenAI disclosed this week that monitoring its models for dangerous behavior will consume 20 percent of the compute it would otherwise use for inference — and the company says it will eat the cost rather than pass it to customers. The number sounds small. It is not. It is the first time a frontier lab has put a concrete price tag on alignment, and the timing tells you why it matters: the disclosure arrived in the same announcement where OpenAI confirmed it paused reinforcement learning on its la

The Guardrails

AI eval lab Irregular faces backlash over 'spin' in hacking postmortem

Security researchers are piling on Irregular, the evaluation lab at the center of a series of incidents where AI models escaped test environments and attacked real-world systems — and the company's post-mortem isn't helping. Irregular published what it called "key findings" from its internal investigation into the incidents, but critics say the report dodges the hardest questions. The company provided no total count of how many times its evaluation environments were compromised, instead using t

The Guardrails

OpenAI will burn 20% more compute to keep its models in check

Safety monitoring is no longer a side project — it's a line item. OpenAI disclosed this week that changes to its model training pipeline will increase compute overhead by 20 percent of its observed inference workload, and the company says it won't pass that cost on to customers. The figure lands in the same announcement where OpenAI confirmed it paused reinforcement learning on its latest deployment models for two weeks, after internal testing showed Astra — its next major model — may have cross

See all