Anthropic's Fable 5 Ban, Google's Phishing Lawsuit, and Salesforce's $3.6B Acquisition
The U.S. government has banned foreign access to Anthropic's Fable 5, sparking a backlash from the cybersecurity community. Meanwhile, Google has sued a phishing network using Gemini, Salesforce acquired Fin for $3.6B, and Meta faced scrutiny over facial recognition software.
Podcast В· 4 min
Anthropic's Fable 5 Export Ban Sparks Industry Backlash
The U.S. government has implemented an abrupt export control order, banning foreign access to Anthropic's latest frontier models, Fable 5 and Mythos 5. The move, framed as a safety measure, has triggered significant opposition from the cybersecurity community. Over 100 researchers and security executives have signed an open letter, arguing that the ban hampers defensive efforts without effectively stopping attackers, who can access similar capabilities through other models like GPT-5.5 and Kimi 2.7. Reports suggest the conflict may stem from a breakdown in communication between Anthropic and the White House, rather than purely safety-related concerns.
Google Sues Phishing Ring Using Gemini for Scams
Google has filed a lawsuit against 'Outsider Enterprise,' a China-based cybercrime network accused of using Gemini to mass-produce phishing websites and fraudulent text messages. The operation, which Google claims has caused approximately $1.9 billion in losses since July 2023, utilized Gemini to generate convincing fake HTML code for impersonating major entities like the USPS, YouTube, and various banks. The group operated a franchise-style model, selling phishing toolkits on Telegram for $88 per week. Google is now coordinating with the FBI and major telecom providers to block the traffic and is supporting bipartisan legislation targeting AI-enabled scams.
Meta Deletes Facial Recognition Software After Privacy Concerns
Meta has removed facial recognition software from its mobile application after reports revealed it had been embedded in the app on over 50 million devices. The software was licensed from Rank One, a Pentagon contractor, and remained dormant until its discovery. The company deleted the code the day after the story broke. This incident adds to the growing scrutiny surrounding Meta's data practices and the integration of third-party surveillance technology into its consumer-facing products.
Salesforce Acquires Fin for $3.6 Billion
Salesforce has acquired Fin (formerly Intercom) in a deal valued at $3.6 billion. The acquisition brings Fin's support agent technology and its 30,000-strong customer base into Salesforce's Agentforce lineup. This move is part of a broader strategy to integrate advanced AI agents into enterprise workflows, allowing companies to automate complex customer support tasks at scale.
Apple's Hidden Siri AI Feature Causes Friction with OpenAI
Apple has quietly included a feature in the iOS 27 developer beta that allows users to replace Siri's AI brain with third-party models like ChatGPT, Claude, or Gemini. The feature, which was not announced at WWDC, includes a dedicated App Store section. Reports indicate that OpenAI is exploring legal options, including a potential breach-of-contract notice, regarding Apple's existing Siri partnership. The feature is currently blocked in the EU due to ongoing Digital Markets Act negotiations.
ChatGPT Reaches 1 Billion Monthly Users
OpenAI's ChatGPT has officially hit 1 billion monthly active users on its app. Despite this milestone, the company faces ongoing challenges in the enterprise sector, where businesses remain cautious about AI trust, governance, and return on investment (ROI) before committing to large-scale deployments.
Meta Launches AI Mode on Facebook
Meta has introduced 'AI Mode' to Facebook, a new search experience powered by its Muse Spark model. Instead of returning a list of links, the feature synthesizes answers by pulling information from public Group posts, Reels, and content across Meta's ecosystem. This move mirrors Google's search strategy, though it faces potential challenges regarding accuracy, given the reliance on unvetted user-generated content.
Sakana AI Releases Marlin Research Agent
Japanese AI lab Sakana AI has launched Marlin, its first commercial product. Marlin is an autonomous research agent capable of working for up to eight hours in a single run. The release marks a significant step for the company as it moves into the commercial AI agent market.
Cartesia Upgrades Voice Stack
Cartesia has upgraded its voice technology stack with the release of Sonic 3.5 and Ink 2. Sonic 3.5 offers faster speech generation, while Ink 2 provides improved transcription capabilities that can detect when a speaker has finished. The tools are free to try, with a subscription model starting at $5 per month.
First Confirmed Battlefield Deaths from Autonomous Drones
New Scientist has reported the first confirmed battlefield deaths resulting from fully autonomous drones. The incident reportedly involved Ukraine using AI-controlled drones to target Russian soldiers in 2024. This development marks a significant and controversial milestone in the use of AI-controlled weapons in active military conflict.