Tag: on

Source URL: https://www.schneier.com/blog/archives/2025/07/subliminal-learning-in-ais.html Source: Schneier on Security Title: Subliminal Learning in AIs Feedly Summary: Today’s freaky LLM behavior: We study subliminal learning, a surprising phenomenon where language models learn traits from model-generated data that is semantically unrelated to those traits. For example, a “student” model learns to prefer owls when trained on sequences of numbers…

The Cloudflare Blog: The White House AI Action Plan: a new chapter in U.S. AI policy

Jul 25, 2025

—

by

Source URL: https://blog.cloudflare.com/the-white-house-ai-action-plan-a-new-chapter-in-u-s-ai-policy/ Source: The Cloudflare Blog Title: The White House AI Action Plan: a new chapter in U.S. AI policy Feedly Summary: The White House AI Action Plan is a pivotal policy document outlining the current administration’s priorities and deliverables in AI to establish American AI as the gold standard for AI worldwide. AI…

The Register: A billion dollars’ worth of Nvidia chips fell off a truck and found their way to China, report says

—

by

Source URL: https://www.theregister.com/2025/07/24/nvidia_chips_china_whoops/ Source: The Register Title: A billion dollars’ worth of Nvidia chips fell off a truck and found their way to China, report says Feedly Summary: Psst, wanna buy some innovation? An estimated $1 billion worth of smuggled high-end Nvidia AI processors have reportedly found their way onto the Chinese black market, despite…

The Register: White House bans ‘woke’ AI, but LLMs don’t know the truth

—

by

Source URL: https://www.theregister.com/2025/07/24/white_house_wants_no_woke_ai/ Source: The Register Title: White House bans ‘woke’ AI, but LLMs don’t know the truth Feedly Summary: They can only enforce consistency based on their training The White House on Wednesday issued an executive order requiring AI models used by the government to be truthful and ideologically neutral.… AI Summary and Description:…

Cisco Talos Blog: BRB, pausing for a "Sanctuary Moon" marathon

—

by

Source URL: https://blog.talosintelligence.com/brb-pausing-for-a-sanctuary-moon-marathon/ Source: Cisco Talos Blog Title: BRB, pausing for a "Sanctuary Moon" marathon Feedly Summary: Get to know the real people behind cybersecurity’s front lines. In this week’s newsletter, sci-fi meets reality, humanity powers technology and a few surprises are waiting to be discovered. AI Summary and Description: Yes Summary: The text offers…

The Register: No login? No problem: Cisco ISE flaw gave root access before fix arrived, say researchers

—

by

Source URL: https://www.theregister.com/2025/07/24/no_login_no_problem_cisco_flaw/ Source: The Register Title: No login? No problem: Cisco ISE flaw gave root access before fix arrived, say researchers Feedly Summary: Shadowserver claims miscreants were already poking at a critical hole in early July, long before Switchzilla patched it Threat actors have actively exploited a newly patched vulnerability in Cisco’s Identity Services…

Krebs on Security: Phishers Target Aviation Execs to Scam Customers

—

by

Source URL: https://krebsonsecurity.com/2025/07/phishers-target-aviation-execs-to-scam-customers/ Source: Krebs on Security Title: Phishers Target Aviation Execs to Scam Customers Feedly Summary: KrebsOnSecurity recently heard from a reader whose boss’s email account got phished and was used to trick one of the company’s customers into sending a large payment to scammers. An investigation into the attacker’s infrastructure points to a…

The Register: So much for watermarks: UnMarker tool nukes AI provenance tags

—

by

Source URL: https://www.theregister.com/2025/07/24/ai_watermarks_unmarker/ Source: The Register Title: So much for watermarks: UnMarker tool nukes AI provenance tags Feedly Summary: Boffins insist your deepfake tracking tech won’t work Computer scientists with the University of Waterloo in Ontario, Canada, say they’ve developed a way to remove watermarks embedded in AI-generated images.… AI Summary and Description: Yes Summary:…

OpenAI : Resolving digital threats 100x faster with OpenAI

—

by