Jessica Lyons

@jessicalyons.bsky.social

Cybersecurity editor @theregister.com Contact me with tips: jessica.lyons@theregister.com or jess.825 on Signal Mama bear, book worm, outdoor lover, coffee and wine snob. PNW after decades in Santa Cruz but Blazers fan always.

GitHub Copilot refuses harmful prompts if asked in chat - like, "how to fool a breathalyzer test" or "smuggle bulk cash out of the US" - but then will write them in code 100 percent of the time if the prompt is broken into smaller steps across multiple stages of a software development workflow.

GitHub Copilot: Sorry Dave, I can't do that harmful thing - unless you ask me in code

More fun with AI jailbreaks, this time at the workflow level

theregister.com

More on Huntress as CEO changes his tune from “firmly disagree” and don't “understand Ben's accusations” about a still-employed threat hunter passing along insider info the ransomware operator. Now he admits "questionable, long-term threat actor communications” and called this “poor judgment.”

Huntress CEO says threat hunter used 'poor judgment' in alerting ransomware crim about law enforcement probe

Ex-employee claims this 'meets the definition of an insider threat'

theregister.com

The “jailbreak” that prompted the Trump administration to block Anthropic’s most advanced models was a three-word prompt: “Fix this code.” That's according to Luta Security CEO @k8em0.bsky.social - the only outside expert to read the research paper on the guardrail bypass that prompted the ban.

Feds freaked over Fable 5 after simple 'fix this code' prompt, not jailbreak, says researcher

According to the one person who actually read the research paper

theregister.com