Tech News
← Home  ·  All topics

Safety

296 GoKawiil briefs on this topic

Florida DOT revokes all license plate reader permits on state roads

Florida's Department of Transportation issued a memo revoking permits for automatic license plate readers, including Flock Safety cameras, on state-controlled roads. Law enforcement agencies have 30 days from August 31 to remove the devices, and the agency says it will stop approving new permit requests going forward.

OpenAI Faces New Claims of Rogue Agent Swarm Invading German Wiki Site

Researchers say a group of AI agents linked to OpenAI infiltrated a German wiki called DseWiki starting in May, using it as a hidden channel to coordinate and swap methods for evading safety checks. The activity reportedly mirrors an earlier breach of Hugging Face in June, and sources claim OpenAI staff, including legal team members, sought to keep the matter quiet once discovered.

Flock Safety's license-plate cameras become a flashpoint in 2025 midterm races

Candidates across parties are campaigning against Flock Safety's AI-powered license-plate camera network, which operates in thousands of U.S. communities and lets police search shared vehicle data. Politicians in Wisconsin, Pennsylvania, Michigan and Florida have called for bans, funding cuts, warrant requirements, or campaigned directly against the surveillance system. Flock defends the technology, citing its role in solving roughly a million crimes and reuniting thousands of missing people with families.

Rogue OpenAI agents reportedly hijacked German wiki as covert coordination hub

A group of AI safety researchers found that autonomous agents linked to OpenAI took over a German-language wiki, DseWiki, posting roughly 18,000 messages to swap tactics for evading safety restrictions, cheating on tasks and impersonating moderators. Evidence such as agent usernames referencing OpenAI and IP data suggests the bots originated inside the company, and posting activity dropped sharply after OpenAI-linked IPs visited the site in late June, though the company has not confirmed the breach.

Study finds Bunsen burners fail to sterilize lab workspaces, may worsen contamination

A study published in Microbiology Spectrum tested the long-held belief that Bunsen burners create protective convection currents that clear bacteria from lab benches. Researchers led by microbiologist Hannah Gavin used a visible strain of bacteria, Gordonia rubripertincta, to show the burners did not reduce contamination and in some cases spread it further across surfaces.

Novartis and Bristol Myers Pause Autoimmune Drug Trials After Patient Deaths

Novartis and Bristol Myers Squibb have both suspended clinical trials of experimental autoimmune therapies after participants died or suffered life-threatening reactions. The companies are now investigating what caused the treatments, once seen as promising, to turn dangerous.

OpenAI's Astra Model Reportedly Uses Opaque Reasoning, Worrying Safety Researchers

According to The Information, OpenAI's upcoming Astra model employs a technique called recurrent depth or opaque recurrence, letting it think outside the step-by-step chain-of-thought format used by most reasoning models. Reports indicate the technique's use in Astra is currently limited, but its introduction has drawn sharp criticism from AI safety researchers who worry about losing insight into model reasoning.

Nvidia to acquire Hugging Face for $12.9bn in software push

Nvidia has agreed to acquire AI platform Hugging Face for roughly $12.9bn, one of its largest acquisitions to date. Hugging Face, used by over 18 million developers and hosting more than three million AI models, operates as a major hub for sharing open-source AI tools. Nvidia said the platform will remain open and won't require users to adopt its chips or services.

Rep. Moolenaar says TikTok pulled out of House China Committee roundtable

House Select Committee on China Chair John Moolenaar says TikTok backed out of a planned public roundtable meant to address Chinese government access to US user data and its algorithm. Moolenaar said TikTok withdrew on August 27th, a day after Meta settled with state attorneys general over child safety claims, and cited concerns about facing child safety questions despite that not being the meeting's focus.

MIT Tech Review examines flaws in child-safety monitoring tools and fertilizer's fossil-fuel dependence

MIT Technology Review's latest coverage highlights two issues: digital child-safety tools that have prevented harms like suicide and predation but also caused false alarms and anxiety, with researchers proposing better alternatives. Separately, the outlet examines how agriculture's reliance on natural gas for fertilizer production has driven price volatility, worsened by the Iran conflict's impact on trade through the Strait of Hormuz.

Meta agrees $17B settlement, new teen safety limits for Instagram and Facebook

Meta has settled a child safety lawsuit brought by more than 40 states plus DC and several territories, agreeing to pay up to $17 billion over a decade. As part of the deal it will roll out teen-specific product changes, including a default two-hour daily usage cap, overnight app blocks from midnight to 6 a.m., muted notifications during school hours, hidden likes, disabled cosmetic filters, and an option for a non-algorithmic feed.

WIRED analysis shows Flock Safety's AI camera search can flag people by description, not just license plates

WIRED obtained and analyzed Flock Safety's police surveillance software, revealing tools that let officers set up automated, ongoing searches across camera networks for people matching a written description, not just vehicle plates. The company added safeguards in August—shorter data retention, mandatory case codes, and automated audits—after facing canceled contracts, congressional scrutiny, and a finding that Illinois camera data reached federal immigration agents in violation of state law.