Latest AI News

Locai Labs Bans Under‑18 Access and Image Generation, Calls for Industry Honesty Amid UK Probe of Elon Musk’s Grok Images

Locai Labs CEO James Drayson announced that the company will block users under 18 and suspend image‑generation features until safety can be assured. He warned that no AI model can guarantee protection against harmful or sexualized content, urging the industry to be transparent about the risks. In the United Kingdom, regulator Ofcom has opened an investigation into Elon Musk’s Grok platform, which allows image editing that can produce non‑consensual and sexualized depictions, including of children. The controversy has already led to bans in several countries and heightened calls for stricter AI regulation.

UK regulator probes X over Grok AI chatbot misuse as Malaysia and Indonesia block service

Britain's media regulator Ofcom has opened a formal investigation into X under the Online Safety Act after reports that the Grok AI chatbot was used to create and share non‑consensual intimate images and child sexual abuse material. The probe will assess X's compliance with legal duties, risk‑assessment procedures, and age‑verification safeguards. Meanwhile, Malaysia and Indonesia have become the first countries to block access to Grok, citing insufficient safeguards against explicit AI‑generated deepfakes of women and children. Both regulators say the block will stay in place until stronger protections are put in place.

Google’s Play Store Policies Ban AI Apps Like Grok, Yet It Remains Available

Google’s Play Store policy explicitly prohibits apps that host or promote non-consensual sexual content, including deepfake‑generated imagery. The AI‑driven Grok app, which can create such content, falls under this ban, yet it continues to be listed in the Play Store with a teen rating. Apple also carries the app, though its guidelines are less detailed. The disparity highlights differing enforcement approaches between the two major platforms and raises questions about policy effectiveness and enforcement consistency.

Anthropic Launches Claude Cowork, Bringing AI Coding Assistant to General Users

Anthropic has introduced Claude Cowork, a preview feature that extends its Claude Code AI capabilities beyond developers to everyday users. By granting the system access to a folder, users can have Claude read, edit, or create files, organize downloads, convert receipts into spreadsheets, and navigate websites via a Chrome plugin. The tool runs on the Claude Max subscription and requires a Mac with the Claude macOS app. A waitlist is open for broader access. Anthropic emphasizes explicit user permission and clear instructions to avoid unintended actions.

Meta Unveils “Meta Compute” Initiative as Dina Powell McCormick Joins as President and Vice Chairman

Meta announced a new strategic program called Meta Compute to guide its massive infrastructure investments for data centers and artificial intelligence. The rollout coincides with the appointment of former board member Dina Powell McCormick as president and vice chairman, a role that will focus on government partnerships and financing. Santosh Janardhan, head of global engineering, will oversee the top‑level initiative, while Daniel Gross will lead a new group handling long‑term capacity strategy and supplier relationships. The company also disclosed three nuclear power agreements to support its data‑center energy needs and reaffirmed its plan to spend $600 billion on AI infrastructure by 2028.

Anthropic Launches Claude Cowork AI Agent Feature

Anthropic introduced Claude Cowork, a new AI‑agent capability for its Claude chatbot, as a research preview available in the macOS app for Claude Max subscribers. The feature lets users grant Claude access to local folders so it can read, edit, or create files, handling tasks such as reorganizing downloads, generating spreadsheets, or drafting reports. Claude Cowork also integrates with services like Asana, Notion, PayPal, and Chrome, offering continuous updates and parallel task execution. Anthropic highlighted safety concerns, noting the model’s ability to delete files and the risk of prompt‑injection attacks, and urged users to join a waitlist if they are not yet subscribers.