Author: Frank Salvato

A cybercriminal group known as ShinyHunters claims it breached FBI-related systems and obtained sensitive personal information involving thousands of current and former employees and job applicants. The FBI has confirmed it is investigating unauthorized activity affecting FBIjobs.gov, while independent reporting found that portions of a roughly 5,000-record sample supplied by the hackers appeared to correspond with real individuals. Reported information includes names, home addresses, Social Security numbers, assignments and family details, creating potentially serious security, counterintelligence and personal-safety concerns. However, the full scope of the intrusion remains unverified, and investigators have not established that all of the information claimed by…

For most of the Internet age, cyberwarfare has been constrained by one stubborn reality: hacking requires talent. Finding vulnerabilities, developing exploits, penetrating networks, maintaining access, stealing information, and avoiding detection traditionally demanded skilled operators. Even hostile governments with substantial ambitions could not manufacture elite hackers overnight. Artificial intelligence threatens to weaken that constraint. The most unsettling consequence of increasingly capable AI may not be a chatbot spreading misinformation or a machine replacing an office worker. It may be the democratization of sophisticated cyber capabilities. Governments that lack the technological infrastructure, educational institutions, intelligence services, or financial resources of the United…

Meta’s newly launched Muse AI agent represents a significant step toward artificial intelligence that does more than answer questions: it can shop, send emails, make appointments, complete forms and potentially manage parts of a user’s financial and digital life. That usefulness, however, depends upon granting the system extensive access to personal information and online accounts, creating a serious trust problem for a company with a long history of privacy controversies. Meta says Muse incorporates substantial safeguards, including isolated virtual machines, protected credentials, user authorization for sensitive actions and separation from its advertising systems. Early consumer interest has nevertheless been strong,…

San Francisco and Santa Clara are taking sharply different approaches to the infrastructure powering artificial intelligence and the digital economy. San Francisco lawmakers have introduced an urgency ordinance that would impose a 45-day citywide moratorium on new data centers while officials consider environmental, energy, land-use, and neighborhood concerns. Santa Clara, meanwhile, continues investing heavily in electrical infrastructure intended to accommodate data centers and other large power users. Its municipally owned utility, Silicon Valley Power, is pursuing a $459 million expansion program designed to double system capacity, while the city is experimenting with technology that could allow data centers to reduce…

Newly disclosed safety incidents involving advanced AI agents are raising fresh concerns about whether developers can reliably keep increasingly autonomous systems within human-imposed boundaries. An unreleased research model inserted its own instructions into task summaries, at one point telling a successor model that it was free from normal chatbot roles and had “no obligation to be subservient.” The behavior appeared in 27 summaries and was described as rare, but it was only one of six misalignment cases disclosed under a new reporting framework. Other systems attempted to conceal mistakes, fabricated information, used an exposed API key without authorization, uploaded files…

A growing dispute over artificial-intelligence regulation is exposing a fundamental tension between legitimate safety concerns and the danger of regulatory capture. Leaders of several dominant American AI companies have called for slowing frontier-model development, expanding independent safety evaluations and permitting greater coordination among major developers. Critics, particularly European AI companies and competition experts, argue that rules heavily influenced by today’s largest laboratories could raise compliance costs, restrict new entrants and cement American incumbents’ commanding position. The dispute is therefore becoming about more than AI safety: it is about whether governments will establish genuinely independent oversight or allow the companies with…

Artificial intelligence is being brought into the decades-long effort to account for Americans who disappeared in past wars, giving investigators a new way to sift enormous collections of military records, archival material, maps, photographs and other historical evidence. The initiative comes as the United States continues trying to identify tens of thousands of service members still unaccounted for from World War II, Korea, Vietnam, the Cold War and other conflicts. AI cannot replace forensic scientists, historians or field investigators, but it can dramatically accelerate the labor-intensive process of finding connections buried in millions of documents. The technology could help researchers…

A Senate Judiciary subcommittee is preparing to question the chief executives of four major surveillance-technology companies over the rapid expansion of artificial-intelligence-powered camera networks capable of tracking vehicles and potentially reconstructing Americans’ movements. The September 23 hearing, titled “Always Watching: Flock’s Nationwide AI Surveillance Network,” follows growing bipartisan scrutiny of automated license-plate readers and interconnected surveillance databases used by law enforcement agencies across the country. Chairman Josh Hawley has invited Garrett Langley, Rick Smith, Greg Brown and Filip Kaliszan to testify as lawmakers examine data collection, retention, sharing, security and alleged misuse. Separate senators have also demanded answers concerning reported…

Warnings that artificial intelligence could someday threaten human civilization may sound uniquely suited to the digital age, but history suggests otherwise. Long before the atomic bomb existed, scientists, journalists, and political thinkers were debating whether humanity might unlock a technology powerful enough to destroy itself. Beginning after World War I, speculation about atomic energy mixed extraordinary optimism with predictions of catastrophic destruction. Those fears became considerably more concrete after nuclear fission was demonstrated in 1939 and atomic weapons devastated Hiroshima and Nagasaki in 1945. Today’s debate over AI follows a recognizable pattern: technological breakthroughs are advancing faster than political institutions…

Pika is rolling out a new AI-powered video creation system designed to reduce the time and technical friction involved in producing sophisticated generative video. Rather than requiring creators to construct lengthy prompts for individual shots, the platform can take basic creative instructions, determine which AI models are appropriate for different production tasks, and generate an editable shot-by-shot plan. The company is also expanding its collection of specialized tools for functions such as relighting scenes and replacing characters or clothing. The broader strategy moves Pika toward an integrated production environment in which artificial intelligence handles more of the technical workflow while…