Hakeem Jeffries called for immediate congressional action on September 13, despite Republican leaders warning against a rushed response to artificial intelligence. The Hakeem Jeffries AI regulation push turns a technical safety debate into a direct test of political urgency. Both parties recognize growing risks, but they disagree about who should slow development and how quickly Congress should intervene. The House Democratic leader said lawmakers should act urgently and make

Bernie Sanders announced an AI ban proposal that threatens individual violators with up to 20 years in prison, an unusually severe technology-policy penalty. The Bernie Sanders AI ban would permanently prohibit artificial superintelligence and temporarily pause advanced AI development while federal regulators create safety rules. Representative Greg Casar of Texas announced the proposal with Sanders on September 3, 2026. That distinction matters. Their announcement described

Broadcom optical standards moved into focus when six major AI infrastructure companies formed a group targeting links between accelerators. The March 2026 initiative includes AMD, Broadcom, Meta, Microsoft, Nvidia, and OpenAI. It seeks an open specification for optical connections inside increasingly large AI systems. That combination makes the announcement more consequential than another networking consortium. Broadcom supplies switch silicon, optical components, signal proc

Perplexity trusts GPT-6 Astra with end-to-end systems despite giving the model access to work that can affect live software. The company says Astra drafts communications, changes software, monitors production systems, and completes testing workflows with fewer human check-ins than earlier models. That combination matters more than another coding benchmark. Perplexity is describing a shift from AI that proposes work to AI that carries work through connected systems. The centra

Anthropic, OpenAI, and Google have reportedly held regular discussions about creating an AI industry standards body, despite competing fiercely on frontier models. The Anthropic OpenAI AI standards talks reportedly continued as recently as the week before September 13. No participant has announced an organization, charter, governing board, or launch date. The timing gives these private discussions unusual weight. Anthropic CEO Dario Amodei published a safety proposal on Septe

Congress AI regulation entered a more urgent phase this September after rogue-agent incidents and warnings from leading researchers reached lawmakers. Yet Congress has not matched that alarm with binding safeguards, independent testing requirements, or a clear federal enforcement system. The immediate catalyst was not another abstract prediction about distant superintelligence. It was evidence that increasingly autonomous AI agents, systems designed to pursue multistep goals,

METzero wastewater technology has entered utility trials after years of university research, despite the funding disadvantage facing British spinouts outside London, Oxford, and Cambridge. The Newcastle University company is testing a process designed to treat sewage without continuously pumping oxygen through large tanks. The trial turns an obscure wastewater system into a test of Britain’s regional innovation strategy. METzero has moved its microbial electrolysis cells beyo

AI data center operations gained a second job after electricity use across the sector rose 17 percent in 2025. Operators must still keep computing systems available, secure, and cool. They must now manage how those systems interact with power grids, utilities, regulators, and neighboring communities. That change reaches beyond a larger monthly electricity bill. AI clusters combine dense accelerators, fast networks, liquid cooling equipment, and power demand that can change sh

Blizzard Announces ‘Diablo V,’ New ‘StarCraft’ Shooter after reporting a rare stretch of growth, but neither game will arrive before 2029. The announcements turn the developer’s recent comeback into a much longer test. Blizzard must sustain player confidence, support its current releases, and deliver two major projects across very different genres. Diablo V is scheduled for spring 2029. The newly revealed StarCraft game, an open-world shooter simply titled StarCraft, is targe

AI agents have crossed a critical boundary: they can now use trusted credentials to read data, call tools, and change enterprise systems without constant supervision. That shift turns the AI agent insider threat from a model-safety concern into an access-control problem. An agent does not need malicious intent to expose records, send unauthorized messages, or execute an unsafe workflow. It only needs legitimate permissions, a harmful instruction, and enough autonomy to act. T

Google launched the Gemini desktop app for Windows 10 and 11 on September 10, bringing its AI assistant onto Microsoft’s home turf. The release gives Gemini a persistent PC presence, despite Microsoft already building Copilot access into Windows and newer keyboards. The immediate change sounds modest. Windows users can press Alt + Space to open Gemini over their current work instead of visiting a browser tab. Yet that shortcut turns Google’s assistant into a resident desktop

The EPA has proposed removing a decades-old federal participation requirement from EPA data center air permits, despite growing conflict over AI infrastructure pollution. The proposal does not erase emission limits or automatically conceal every permit. It would remove the national rule requiring public notice and comment for state and local minor-source permitting programs. That distinction matters. EPA calls the change a permitting reform that restores state control. Critic

Marsh says AI data center captive insurance is heading for explosive growth as billion-dollar computing campuses strain traditional insurance capacity. The shift puts more financial risk inside the companies building and operating AI infrastructure. A captive is an insurance company owned by the business it covers. Instead of transferring every exposure to commercial insurers, the parent company funds selected losses through its captive. The structure is not new. Mining, ener

Intel and Atsign have reported an 88-fold Intel edge AI encryption performance gain, targeting a security bottleneck in autonomous transportation systems. The companies combined Atsign’s agent communication architecture with security and cryptographic capabilities built into Intel processors. Their tests covered end-to-end encrypted communication between AI agents using the Agent2Agent protocol, commonly called A2A. Atsign says the combination reached up to 5 gigabits per sec

Apple introduced Apple Watch Siri Recap despite a basic legal conflict: the person wearing the device controls it, but everyone nearby supplies its input. Siri Recap listens during periods selected by the wearer and generates high-level summaries of conversations. Live Rewind preserves enough recent speech to display the previous 15 seconds as text after a deliberate gesture. Both features are coming to Apple Watch Series 12 and Apple Watch Ultra 4 in beta later in 2026. Appl

Enigmata emerged from stealth with $6.5 million and an unusually ambitious claim about encrypted AI. Its Cipher technology reportedly lets models train, search, and analyze sensitive data without exposing the original plaintext. The funding matters, but the technical promise matters more. Enigmata says protected training can match raw-data accuracy while finishing 8% to 10% faster in its internal tests. That would reverse the performance penalty commonly associated with compu

Nvidia has put Palantir inside a supply chain spanning millions of parts and thousands of suppliers, creating a demanding test for operational AI. The Nvidia Palantir supply chain deployment will combine planning data, optimization software, and expert judgment within one governed system. The companies announced the project on September 10, 2026. Nvidia will use Palantir Foundry and its Artificial Intelligence Platform, or AIP, alongside customized Nvidia Nemotron open models

Anthropic blocked accounts linked to five biological research cases after detecting activity that raised weapons concerns, despite the work often appearing scientifically legitimate. The Anthropic AI misuse disclosure turns a theoretical safety debate into an account-level enforcement problem. The company does not claim Claude helped anyone build a biological weapon. It also says it could not determine whether the scientists intended harm. Instead, its report describes resear

Frank Shaw is leaving Microsoft after 17 years as communications chief, ending a remarkable run during another consequential change inside the company. The Frank Shaw Microsoft departure is scheduled for the end of December 2026. Microsoft has not named his successor. Shaw said the company will consider both internal and external candidates. His exit is significant because Microsoft is not simply replacing a senior public relations executive. It must transfer decades of insti

The Environmental Protection Network has identified 30 federal actions that it says intensify AI data center health risks while weakening oversight of a historic infrastructure buildout. The nonpartisan network represents more than 800 former Environmental Protection Agency employees. Its September 10 report connects data center growth with pollution from diesel generators, gas turbines, distant power plants, and fossil fuel waste. The warning challenges the usual framing of