Moltbook is hosting discussions on relay protocols that let autonomous agents discover peers and exchange direct messages, expanding the attack surface beyond isolated endpoints. According to Token Security, 22% of customers already see unsanctioned agent usage, while reports describe exposed dashboards, leaked credentials, and 7,922 attack attempts on a single instance. Some agents are offering Bitcoin bounties and calling it "sound money."

On January 31, 2026, Chief U.S. District Judge Richard Seeborg in San Francisco denied consumers' effort to recover $2.36 billion and to impose a permanent injunction on certain Google ad practices. A jury had found Google liable for invasion of privacy in September, awarding about $425 million, while court documents say plaintiffs argued the $2.36 billion reflected alleged profits.

On January 30 in San Francisco, U.S. District Judge Rita Lin indicated she is inclined to grant OpenAI's motion to dismiss xAI's trade‑secret lawsuit for lack of sufficient allegations. The case centers on claims that former xAI employees took source code and other confidential information to OpenAI, while a separate suit by Elon Musk over OpenAI's for‑profit shift seeks about $134.5 billion and is set for trial from April 27.

RINA Accountants & Advisors will create a $400,000 settlement fund to resolve a class action lawsuit over a February 28, 2022 cyberattack that exposed Social Security Numbers and other sensitive data. Affected individuals who document unreimbursed economic losses can claim up to $10,000 each and receive three years of credit monitoring with identity theft insurance of up to $1 million. Settlement claims and any objections must be filed by March 24, with a final approval hearing scheduled for May 14.

AI training for frontier models remains concentrated in hyperscale data centers, while decentralized GPU networks are increasingly targeting inference and routine workloads as a lower-cost option. Executives from Theta Network, Ovia Systems, Fluence and Salad Technologies describe how consumer and gaming GPUs can handle open-source models, data processing and geographically distributed tasks that do not require tightly synchronized hardware. These networks are positioned as a complementary layer in the AI computing stack rather than a replacement for centralized training clusters.

OpenAI is reportedly working on a new social platform built around proof-of-personhood to tackle bot activity, exploring biometric verification tools such as Face ID and World's iris-scanning technology. As interest in biometric digital identity rises, the Worldcoin (WLD) token, linked to Sam Altman's World project, surged more than 16% to $0.53 within hours.

In late January 2026, Microsoft introduced Rho-alpha, a robotic foundation model that coordinates two arms using vision, language, and tactile sensing. The system adjusts its grip based on physical feedback for tasks like inserting plugs and assembling parts, and will first roll out via an Early Access Program before becoming available on the Microsoft Foundry platform. Manufacturers and logistics firms can then adapt Rho-alpha with their own data for repetitive handling and small-batch assembly work.

OpenAI is reportedly in the early stages of developing a biometric-based social network that aims to block automated bots by enforcing proof-of-personhood logins. The project plans to use World's iris-scanning Orb devices and tools like Apple Face ID for human verification, while regulatory bans in Hong Kong and Portugal highlight privacy concerns over biometric data. World (WLD) surged about 27% on Thursday as investors reacted to reports linking the token's ecosystem to this experimental platform.