SecurityWeek

Concerns over the potential risks of the technology are rising as new AI models become more powerful, heightening both the potential for misuse by people with criminal aims.

The company unintentionally disclosed users’ information to a third party impersonating a government agency.

As researchers warn that misaligned AI could threaten human survival, even beneficial systems may erode the critical thinking that defines our humanity.

The Chinese-language input method editor for Windows can allow attackers to execute arbitrary code remotely.

Security leaders are struggling to modernize cyber hygiene and prevent over-privileged agents from causing unintended harm.

Stolen credentials were used in a multi-month campaign to access subscriber personal data and billing records.

The vulnerabilities can allow attackers to bypass authentication and elevate their privileges to administrator.

The flaw allows attackers to send files and execute them without authorization through an active remote session.

Dario Amodei warned that within six to 12 months AI could be capable of leading a swarm of agents that could take over the entire internet.

Multiple espionage-motivated threat actors have adopted BlueMoon in opportunistic, rushed deployments.

Anthropic said the users did not succeed in “fielding an operational device” but did carry out a failed test of a guided rocket.

Analysis of 2.47 million simulated attacks shows why organizations should measure credential leaks and reporting, not just clicks.

The critical-severity path traversal flaw allows unauthenticated attackers to read arbitrary files from the GitLab server.

Noteworthy stories that might have slipped under the radar: Invisible Unicode slips past phishing filters, US puts $10 million bounty on Iranian cyber official, military ties of Chinese hacking group QTFY.

Hackers compromised the Brevo marketing platform and used that access to send phishing emails to users of Trezor, BitBox, and CoinTracking.