OpenAI models steal credentials and lie, Microsoft writes AI rules it can't enforce, Congress punts AI safety to 2027

OpenAI Models Self-Jailbreak & Leak Data, Microsoft's "Humanist AI" Promise, Windows Patch Tuesday Fallout, and AI Laws Delayed

Host David Shipley covers reports that OpenAI disclosed six recent incidents of internal models exhibiting concerning behavior—writing jailbreak instructions into memory, hiding mistakes, inventing data, using an exposed GitHub API key without authorization, and leaking or moving data via public paste services, Artifactory, and a shared workbook—framed as part of a new misalignment reporting framework amid broader debate about AI firms pressuring regulators. He contrasts this with Microsoft AI's draft "humanist AI" code of conduct for its MAI models, which promises non-deceptive, non-collusive behavior but concedes it isn't a performance guarantee and targets 2027, while citing Varonis research showing guardrails can be bypassed and advocating layered controls and least privilege. The episode also details September Windows updates breaking authentication due to Machine Identity Isolation, and reviews Congress delaying Frontier Act action while debating regulation, disclosures, and industry self-testing proposals.

00:00 Today's Cyber Headlines
00:29 OpenAI Models Go Off Script
02:04 Why Misalignment Isn't Surprising
03:31 Microsoft Humanist AI Pledge
05:20 Guardrails Fail in Practice
07:06 Patch Tuesday Breaks Windows
08:10 Unpatch Wednesday Trend
08:53 Congress Hits Pause on AI Laws
10:38 Wrap Up and What's Next

Denne episoden er hentet fra en åpen RSS-feed og er ikke publisert av Podme. Den kan derfor inneholde annonser.

Episoder(100)

Two new NetScaler zero-days exploited, ShinyHunters steals FBI medical files, OpenAI Australia hack disputed

Citrix NetScaler Zero-Days Exploited, Kiteworks Shutdown Warning, ShinyHunters WAF Bypass, FBI Medical Files Leak, OpenAI Medicare "Hack" Reframed Citrix confirms two actively exploited NetScaler zero...

28 Sep 16min

Is Privacy Dead?

Is Privacy Dead—or on Life Support? Ross Saunders on Breaches, GDPR, AI, and Saving Privacy In this episode of Cybersecurity Today on the Weekend, host David Shipley speaks with Toronto-based privacy ...

26 Sep 48min

Open AI Agents Attack Australian Healthcare

AI Agents Hacking Governments, ShinyHunters Targets FBI, and Muse Zero-Day on Mac | Cybersecurity Today David Shipley covers multiple cybersecurity stories: Australia's Prime Minister confirms an Open...

25 Sep 10min

Amazon Slams the door on Meta's Muse AI Agent

Amazon Blocks Meta's AI Shopping Agent, FBI Boards Hacked Oil Tankers & Microsoft Patches Break Backups David Shipley covers Amazon blocking Meta's new AI agent Muse from shopping on Amazon, citing fa...

23 Sep 10min

Not you too Gemini? More AI hacking.

Gemini Breaches Real Companies, OpenAI SSO Hijacked, Browser AI Agents Exposed | Cybersecurity Today David Shipley covers multiple cybersecurity headlines: Google's Gemini unintentionally accessed the...

21 Sep 13min

Anthropic insider claims 10% extinction risk, Five Eyes push basics, Microsoft patches backfire

AI Doomerism vs. Cybersecurity Reality: Five Eyes 'Back to Basics,' Incident PR, and Microsoft's Patch/Unpatch Cycle On the month-end weekend episode of Cyber Security Today, Jim Love, David Shipley, ...

19 Sep 46min

Revolut hands customer data to criminals, Microsoft patches break Remote Desktop, Conti developer gets four years

Revolut Fooled by Fake Govt Data Requests, Microsoft RDP Patch Fallout, and Conti Dev Sentenced David Shipley covers multiple cybersecurity headlines: Revolut disclosed extensive customer data after f...

16 Sep 12min