Get tomorrow's brief in your inbox
Today: The FCC is proposing rules requiring telecoms to collect government IDs and physical addresses from all wireless customers, eliminating anonymous prepaid phone purchases under the guise of stopping robocalls (proceeding 13-97, public comment period closes June 25). The US government forced Anthropic to disable two advanced AI models citing national security export controls after claiming knowledge of a jailbreak technique, marking the first use of export control authorities against deployed AI models rather than hardware. Belarus-linked hackers expanded phishing operations to target Gmail accounts of Polish officials and their families, shifting from work email to personal accounts since March.
FCC Proposes Elimination of Anonymous Prepaid Phone Service
The FCC has opened proceeding 13-97 proposing to require all US wireless carriers to collect government-issued identification numbers and physical addresses from every new and renewing wireless customer. The proposal frames the data collection as a robocall prevention measure but explicitly states the collected data could be used for a range of other law enforcement purposes. The rule would eliminate the ability to purchase prepaid phones without identity verification, affecting abuse victims, journalists, whistleblowers, and refugees who rely on anonymous communications. Privacy advocates note the proposal is unlikely to stop actual criminal robocall operations, as fraudsters can easily create fake documentation, but will expose vulnerable populations to surveillance risks. The public comment period closes June 25, 2026.
US Government Forces Anthropic to Disable Advanced AI Models Under Export Controls
The US government issued an unpublished directive on June 13, 2026, invoking national security export control authorities to force Anthropic to immediately disable its Fable 5 and Mythos 5 AI models for all customers worldwide. The directive bars any foreign national from accessing the models, including foreign nationals within the United States and Anthropic's own employees. The government provided only verbal evidence of a claimed jailbreak technique, which Anthropic reviewed and found to be a minor, previously known vulnerability also present in publicly available models like OpenAI's GPT-5.5. This marks the first known use of export control authorities to recall a deployed commercial AI model rather than restricting chips or hardware. Anthropic is complying while disputing the action's basis, warning that applying this standard industry-wide would halt all new frontier model deployments. The move follows February's designation of Anthropic as a "supply chain risk" by Defense Secretary Pete Hegseth after military contract negotiations failed.
Belarus-Linked GhostWriter Campaign Targets Personal Gmail Accounts of Polish Officials
CERT Polska reported that the Belarus-linked GhostWriter threat group (also tracked as UNC1151 and Storm-0257) has shifted tactics since March 2026 to target personal Gmail accounts of Polish government officials, researchers, journalists, law enforcement personnel, and their family members. The group previously focused on work email accounts and Polish email providers. The phishing campaigns steal login credentials and two-factor authentication codes to access email accounts, contact lists, and linked social media profiles. CERT Polska observes new phishing domains deployed almost daily. The attackers sometimes guess likely Gmail addresses when they lack precise targets, resulting in misdirected phishing attempts to unrelated individuals with similar names. GhostWriter has been linked to Belarusian state intelligence and conducts both credential theft and disinformation operations aimed at undermining Poland's relationships with Ukraine, the US, and NATO.