AI-curated news for your industry

CISA released a report on Thursday disclosing that a contractor leaked sensitive information by uploading build and deployment code containing administrator credentials and AWS access keys to a personal GitHub repository. The incident was brought to light by independent journalist Brian Krebs in mid-May, prompting CISA to take immediate response measures, confirming no unauthorized use or data exposure, and committing to strengthen key management and reporting channel development.

According to filings disclosed by multiple U.S. state regulators, Atlanta-based auto and renters insurance provider AssuranceAmerica experienced a cyberattack in March this year, leading to a data breach affecting over 6.9 million customers, involving sensitive information such as names, policy numbers, driver's license numbers, and Social Security numbers. The attack originated from a targeted intrusion on an employee, and the company has taken emergency measures and reported the incident to law enforcement.

U.S. companies are integrating cybersecurity risk into enterprise risk management frameworks, and driven by AI adoption and business resilience, cyber risk has become a core issue for boards and senior management.

A recent Cisco report indicates that as AI applications place higher demands on enterprise network infrastructure, IT executives are deeply worried about the resulting cybersecurity consequences. Respondents identified security complexity as the biggest challenge brought by AI-driven network demands and considered security concerns a direct barrier to AI scaling. Based on a survey of 3,472 CIOs and technology leaders worldwide conducted from March to April 2026, the report highlights key issues such as an expanded attack surface, shadow AI, inconsistent policy enforcement, and limited visibility.

Security researchers have been tracking attack campaigns targeting Cisco SD-WAN vulnerabilities since early 2026, which have impacted critical government websites and infrastructure providers. Experts note that edge infrastructure is the highest-value target in enterprise security, and Cisco products, due to their widespread deployment in government and enterprise networks, are frequently targeted by attackers.

Bitdefender's annual report reveals that 55% of cybersecurity professionals have been asked to conceal security incidents, a proportion that rose during 2024-2025 and then stabilized; meanwhile, over half of respondents believe AI favors attackers. The report also highlights issues such as shadow AI and the confidence gap between management and frontline staff.

Anthropic reopened access to its advanced AI models Fable and Mythos on Wednesday, following the Trump administration's lifting of export control restrictions. Fable 5 is available to general users, while the more powerful Mythos 5 is restricted to a trusted partner consortium. The company stated it has improved safety classifiers and is working with government and industry partners to establish more formal review standards.

The U.S. Department of Homeland Security (DHS) on Tuesday proposed a new public-private collaborative framework for critical infrastructure to replace the previous framework abolished shortly after President Trump returned to the White House. The new framework, named the "Alliance for National Community Harmony and Operational Resilience - Critical Infrastructure National Council" (ANCHOR-CI), will serve as an umbrella structure for multiple advisory committees, designed to foster cooperation between the government and the private sector on cybersecurity and critical infrastructure resilience. It encompasses four types of committees—sector-specific, cross-sector, industry, and regional—managed by the Cybersecurity and Infrastructure Security Agency (CISA) with an initial term of two years. However, the new framework does not include the liability protections under the former CIPAC framework, raising concerns within the industry.

OpenAI announced on Friday that, at the request of the Trump administration, it will restrict the public release of its new GPT-5.6 series models (Sol, Terra, Luna), offering limited previews to a small number of trusted partners first. The company stated it will temporarily comply with the government's request and plans a broader release within weeks. Previously, the government had imposed export controls on Anthropic's models, and this incident highlights a shift in AI regulatory policy.