Author: Frank Salvato

Thirty new lawsuits filed by survivors of the February 10 mass shooting in Tumbler Ridge, British Columbia, have dramatically expanded the legal challenge facing OpenAI and CEO Sam Altman, bringing the total number of related claims to 37. The plaintiffs—including students, teachers and a school principal—allege that OpenAI had identified shooter Jesse Van Rootselaar as a potential threat months before the attack after automated systems flagged violent ChatGPT conversations and company personnel reviewed the account. OpenAI banned the account but did not notify law enforcement, saying the activity failed to meet its threshold for an imminent and credible threat. The…

Burbank’s historic entertainment infrastructure is increasingly being occupied by a new generation of creator-led production companies as conventional Hollywood filming continues to struggle. Digital producers including Smosh, Mythical Entertainment, Alan Chikin Chow and Dhar Mann Studios are operating increasingly sophisticated facilities, employing sizable staffs and attracting partnerships with major entertainment companies. Smosh moved into a 32,000-square-foot facility this spring, while digital-production shoot days increased 47% year over year during the second quarter to 661, even as overall Los Angeles-area on-location production declined. The transformation suggests something larger than influencers simply graduating to soundstages: an entrepreneurial entertainment sector built around direct…

A Chinese technology conglomerate blacklisted by Washington over national-security concerns continued gaining access to some of America’s most advanced artificial-intelligence hardware through subsidiaries, intermediaries and overseas data centers. Its Silicon Valley-linked operation exported at least $5.6 billion in advanced technology to Southeast Asia between April 2024 and February 2026, including more than $3 billion in computers equipped with cutting-edge Blackwell chips. The case exposes a serious weakness in U.S. export controls: restrictions aimed at a named Chinese parent company can be circumvented through affiliates, newly organized entities and foreign computing infrastructure, even as American officials insist advanced chips must be…

Thousands of authors awaiting compensation from a $1.5 billion copyright settlement with artificial-intelligence developer Anthropic are now confronting an unexpected second fight: disputes with publishers, agents, and other parties over who is entitled to the money. The settlement covers more than 482,000 books Anthropic obtained from pirate libraries while assembling material to train its Claude AI models, with approximately $3,000 allocated per eligible work. Although the settlement was designed to compensate copyright holders, competing claims have emerged over books whose publishing rights reverted to authors years ago, works involving multiple rights holders, and contracts containing different royalty or ownership arrangements.…

A cybersecurity research team demonstrated how rapidly artificial intelligence is lowering the technical barriers to sophisticated cyberattacks by developing WeWorm, a self-propagating “zero-click” exploit targeting WeChat users on both iOS and Android devices. The attack exploited a memory-corruption vulnerability in WeChat’s internet-calling system and could compromise an account simply through an incoming call from an account already on the victim’s friend list; the victim did not need to answer or click anything. Once compromised, the account could automatically call its contacts and continue spreading. Researchers said AI helped discover the vulnerability and produce a remote-code-execution exploit in roughly two days,…

An experimental drug designed with the help of artificial intelligence has produced early evidence that it may do more than treat idiopathic pulmonary fibrosis: it may influence biological aging itself. Rentosertib, developed by using AI to identify the TNIK protein as a target and generative AI to design a molecule capable of inhibiting it, was evaluated through blood samples collected during a randomized Phase 2a clinical trial. Researchers analyzed 2,841 proteins from 42 participants using six independently developed proteomic “aging clocks,” and all six registered movement toward younger predicted biological ages among treated patients. The strongest signal appeared in patients…

A senior AI alignment researcher has placed the probability of artificial intelligence killing all humans within the next decade at greater than 10 percent, an extraordinary warning amplified by the resignation of another researcher who accused leading AI developers of racing irresponsibly toward self-improving superintelligence. The concern is not centered on today’s publicly available models, which the researcher says present relatively low immediate risk, but on future systems capable of recursively improving themselves until their abilities exceed human control. Most troubling is the acknowledgment that researchers working directly on AI alignment do not yet have a reliable solution for keeping…

San Francisco’s driverless-car market is becoming an increasingly important test of whether competition can prevent autonomous transportation from being dominated by a single operator. Zoox is challenging Waymo’s substantial head start with a radically different purpose-built robotaxi, free introductory rides, aggressive local marketing, festival sponsorships and experiential promotions intended to turn the vehicle itself into a recognizable consumer brand. Waymo remains far ahead in fleet size, geographic reach and commercial operations, but Zoox is betting that its distinctive vehicle and rider-focused experience can create an alternative as autonomous transportation moves from technological novelty toward a mainstream service. Key Takeaways In-Depth…

Meta has launched Muse, a new personal artificial-intelligence agent designed not merely to answer questions but to perform real-world tasks for users, including sending emails, managing calendars, booking travel, shopping, filling out forms, making payments and pursuing longer-term projects. Available initially in the United States through a dedicated app and WhatsApp, Muse represents Meta’s bid to move AI from passive chatbot assistance toward autonomous digital agency. The company says Muse operates inside a dedicated secure virtual machine and gives users control over the applications and information it can access. Yet internal testing reportedly uncovered problems ranging from unexpected data uploads…

For more than a decade, thousands of educated Kenyans built livelihoods writing academic papers and completing assignments for college students in the United States and other wealthy countries. The work existed in an ethically dubious corner of the digital economy, but it paid substantially better than many conventional jobs available to young Kenyan graduates. That arrangement began collapsing after generative artificial intelligence made sophisticated writing available almost instantly and at little or no cost. Writers who once handled dozens of assignments each month have watched orders and earnings disappear, while some surviving workers now use AI themselves or are paid…