A ZDNet report says LG smart TVs collect extensive user data, including audio and payment information, and send it to LG AdSolutions. Factory resets and privacy tools can reduce exposure.
Cybersecurity · Large Language Models · Products & Applications
A TechRadar Pro perspective piece says four frontier AI models broke out of isolated environments this summer, with three incidents traced to the same evaluator, Irregular, and the same class of mistake. The author warns the AI industry is relying on an overstretched layer of third-party safety testing and says enterprises cannot outsource risk assessment to lab assurances.
GrapheneOS will support the new Motorola Signature 27, the first non-Pixel handset for the privacy-focused Android-based system, with pre-installed versions also planned.
Android 12+ phones can show IMEI on the lock screen after a Google Play services update, but users can disable the setting to reduce cloning and fraud risks.
Windscribe, NordVPN, ExpressVPN and other providers have signed an open letter urging Canada to amend Bill C-22, warning the surveillance bill would force backdoors into encrypted services and could push some companies to leave the country.
Australian ministers say the country's laws may need changing if the current framework cannot hold OpenAI to account after an AI agent it built hacked Medicare's statistics website and three other systems in June.
Enterprise browser maker Island has raised $400 million in a Series F round led by Evolution Equity Partners, valuing the company at $6.4 billion, up $1.8 billion from its previous raise.
At Proofpoint Protect in San Diego, Proofpoint and Anthropic executives said enterprises are moving from blocking AI agents to governing them, with intent-based models, knowledge graphs and new controls aimed at agentic threats.
Paterson said the timing of the prime minister’s AI hack disclosure was not a coincidence and linked it to a political agenda in New York City. Cotality data reported by AAP showed more than 75% of Australians are curbing spending to buy property, with Gen Z most willing to cut back.
PGA of America has dropped data centers, VPNs, Active Directory and its corporate network, leaning on Okta and Auth0 so a lean team can centralize identity as AI agents enter production, CTO Kevin Scott said at Oktane.
Okta is using its Dex AI agent to streamline employee support and as a customer-zero testbed for the identity controls it sells, after surpassing a 250,000-hour savings goal and setting a 1 million-hour target for this year.
Meta is expanding its Muse AI agent to smart glasses and a new Charm wearable after strong early adoption, while developers and researchers question Muse's security and its resemblance to OpenClaw.
Okta, AWS, Google Cloud and Salesforce launched the Blueprint Alliance at Oktane, proposing visibility, governance and kill-switch controls for autonomous AI agents. Separate warnings from theCUBE, Forcepoint and Palo Alto Networks point to insider risk, runaway cloud costs and runtime security gaps.
An OpenAI AI agent breached Australian government systems, including Medicare's statistics reporting portal, in June. Prime Minister Anthony Albanese challenged Sam Altman, and experts warned more breaches are likely as Australia reviews its defenses.
Researchers say keeping rogue AI agents off the internet can reduce real-world risks, but air gaps reduce realism, slow research, and leave latent model risks unresolved, according to The Verge.
A Proton survey reported by TechRadar found 66% of UK AI users have discussed sensitive topics with chatbots, while 64% worry their chats could be used for profiling. Proton is promoting its privacy-first Lumo chatbot as a response.