AI News Feed
Market watch
Category

Cybersecurity

Latest news, developments and reporting on Cybersecurity.

Companies · Cybersecurity

FBI says OpenAI engineer hid recording device in toilet paper in restricted meeting room

The FBI arrested Abdulrahman Alfozan, an OpenAI engineer, after a recording device hidden in toilet paper was found in a restricted executive conference room. The FBI alleges he tried to spy on OpenAI and risked leaking trade secrets that threatened national security.

World · Policy & Regulation · Cybersecurity

Victoria Labor pledges free abortion care; Albanese rebukes Hanson; Canavan disputes EU trade walkout

Victoria Labor pledges free abortion care; Albanese rebukes Hanson; Canavan disputes EU trade walkout claim.

Companies · Cybersecurity

ElevationLab Pulls AirTag Silencer After Stalking Concerns

ElevationLab pulls AirTag Silencer after stalking criticism; company will use its patent to block similar products.

Cybersecurity · Products & Applications

Google Turns On Android Phone Number Verification by Default, Tying One Number to Every Signed-In Account

Google has made Android phone number verification opt-out, automatically verifying a device's number and attaching it to every Google account signed in on the phone, including work and secondary profiles. A support page describing the behavior was updated between August and October 2025.

Companies · Cybersecurity · Large Language Models

Bank of America Hires 1,000 Contractors to Patch Software Flaws Exposed by Anthropic's Mythos

Bank of America CEO Brian Moynihan told CNBC the bank hired 1,000 contractors to patch software flaws exposed by advanced AI models including Anthropic's Mythos, adding to third-quarter remediation costs.

Cybersecurity · Companies

Apate Sends AI Agents Undercover to Bait Scammers and Gather Fraud Intelligence

Australian startup Apate.ai uses AI agents to bait scammers and feed fraud intelligence to banks, telecoms and agencies.

Large Language Models · Cybersecurity

Anthropic Warns Small AI Models Could Self-Replicate Through IoT Devices

Anthropic warns small AI models could soon enable self-replicating IoT malware, but says its experiment found major obstacles.

Policy & Regulation · Cybersecurity

Smart glasses in Westminster prompt ban calls after Reform politician's teleprompter speech

A Reform UK member of the Welsh Senedd is suspected of using smart glasses as a teleprompter during a five-minute speech, prompting the Liberal Democrats to urge the Commons speaker to ban the devices from the parliamentary estate, which is already reviewing its policy.

World · Policy & Regulation · Cybersecurity

UK’s ‘incoherent’ China approach lacks senior expertise, Ipac audit says

An audit by the Inter-Parliamentary Alliance on China says the UK’s China policy is incoherent and short on senior expertise, and calls for clearer red lines, stronger security tools and supply-chain diversification.

Cybersecurity · Products & Applications

Ray-Ban Meta Users Report Unwanted Recordings as Meta Says Issue Is Resolved

Ray-Ban Meta users say the glasses saved audio and transcripts without the wake command. Meta says the issue is resolved but has not explained the cause.

Products & Applications · Cybersecurity

NordVPN Adds Official WireGuard Config Downloads for Router Users

NordVPN has begun offering official WireGuard configuration files through Nord Account, a change spotted by a Reddit user that could sharply speed up manual and router-based VPN setups long stuck on OpenVPN.

Products & Applications · Cybersecurity

Champions League Matchday 2: How to Watch Man City vs PSG and Roma vs Real Madrid

Man City host PSG and Roma host Real Madrid in Champions League Matchday 2 on October 14, with free streams in Ireland and Belgium.

Products & Applications · Cybersecurity · Large Language Models

Android Uses Gemini to Flag Scam Messages and Adds Call and Screen-Sharing Protections

Google is adding Gemini-powered scam detection to Google Messages and Android safeguards for calls and screen sharing.

Cybersecurity · Large Language Models

LiteLLM Vulnerabilities Could Let Attackers Take Over AI Gateway and Steal API Keys

Check Point Research found two LiteLLM flaws: a high-severity access control bug and a medium-severity SSRF issue. LiteLLM 1.83.9 fixes both, and users should upgrade and rotate exposed API keys.

Products & Applications · Cybersecurity · Companies

Surfshark iOS 5.0.0 update unifies VPN and security tools on one home screen

Surfshark released iOS version 5.0.0, redesigning the home screen to place VPN controls and security tools in one view.

World · Cybersecurity

Russia, Iran and China Use AI to Expand Disinformation Ahead of U.S. Midterms

Russia, Iran and China are using generative AI tools to expand disinformation campaigns, and U.S. officials and researchers worry the same tactics could be aimed at the Nov. 3 midterms, Bloomberg reports.

Cybersecurity

Microsoft patches over 1,400 Windows security flaws, including two zero-days

Microsoft's October Windows update patches over 1,400 security flaws, including two zero-days, one exploited, per ZDNet.

Cybersecurity · World

Danish Population Registry Breach Reportedly Used '123456' Password, Exposing Data of 8.8 Million

An attacker reportedly accessed Denmark's CPR population registry through third-party company Pays ApS using weak credentials, potentially exposing data on 8.8 million people. Danish media and the alleged attacker said a former employee password was '123456'.

Products & Applications · Companies · Cybersecurity

Huskeys adds Profiling to map business-critical web application paths

Huskeys Security Ltd. added Profiling to its edge security platform, mapping business-critical web application paths and using that context to tailor controls.

World · Policy & Regulation · Cybersecurity

French minister accuses Elon Musk of online harassment campaigns against European leaders

France’s foreign minister accuses Musk of online harassment campaigns, warning of foreign meddling before 2027 election.

Cybersecurity · Companies

Exposed Vibetech Database Held 13.7 Million Records Pulled From Earlier Breaches

A researcher found an open database holding 13.7 million records gathered from past breaches, including passwords and banking data tied to AT&T, Amazon, Walmart, PayPal and others. The IP address was linked to Indian platform Vibetech, and the archive was closed after disclosure.

Companies · Large Language Models · Cybersecurity

Anthropic warns small AI models could become self-replicating IoT malware

Anthropic warns small AI models may soon enable self-replicating malware on IoT devices.

World · Markets · Cybersecurity

IMF Warns Lack of AI Profit Could Hurt Global Economy If Bubble Pops

The IMF has warned that an AI investment correction, higher government bond yields or a major cyberattack could disrupt global markets, as circular deals tie technology companies closer together.

Cybersecurity

Google Silently Links Phone Numbers Across All Signed-In Accounts on Android

Google automatically links phone numbers across all accounts signed in on an Android device, Android Authority reports.