Linux fréttir

An anonymous reader quotes a report from MIT Technology Review: A team at Microsoft says it used artificial intelligence to discover a "zero day" vulnerability in the biosecurity systems used to prevent the misuse of DNA. These screening systems are designed to stop people from purchasing genetic sequences that could be used to create deadly toxins or pathogens. But now researchers led by Microsoft's chief scientist, Eric Horvitz, says they have figured out how to bypass the protections in a way previously unknown to defenders.The team described its work today in the journalScience. Horvitz and his team focused on generative AI algorithms that propose new protein shapes. These types of programs are already fueling the hunt for new drugs at well-funded startups like Generate Biomedicines and Isomorphic Labs, a spinout of Google. The problem is that such systems are potentially "dual use." They can use their training sets to generate both beneficial molecules and harmful ones. Microsoft says it began a "red-teaming" test of AI's dual-use potential in 2023 in order to determine whether "adversarial AI protein design" could help bioterrorists manufacture harmful proteins. The safeguard that Microsoft attacked is what's known as biosecurity screening software. To manufacture a protein, researchers typically need to order a corresponding DNA sequence from a commercial vendor, which they can then install in a cell. Those vendors use screening software to compare incoming orders with known toxins or pathogens. A close match will set off an alert. To design its attack, Microsoft used several generative protein models (including its own, called EvoDiff) to redesign toxins -- changing their structure in a way that let them slip past screening software but was predicted to keep their deadly function intact. "This finding, combined with rapid advances in AI-enabled biological modeling, demonstrates the clear and urgent need for enhanced nucleic acid synthesis screening procedures coupled with a reliable enforcement and verification mechanism," says Dean Ball, a fellow at the Foundation for American Innovation, a think tank in San Francisco.

Read more of this story at Slashdot.

Cool kids drank the aggressive micro-management Kool-Aid

On Call By Friday morning, techies may need a jolt of energy to get through the final day of the working week, so we deliver it in the form of a new instalment of On-Call, the weekly reader-contributed column that shares your tales of trying to deliver speedy tech support.…

alternative_right shares a report from the Smoking Gun: Minutes after vandalizing 17 cars in a Missouri college parking lot, a 19-year-old sophomore had a lengthy ChatGPT conversation during which he confessed to the crime, asked about the possibility of getting caught, and wondered, "is there any way they could know it was me," according to a police probable cause statement. Ryan Schaefer was arrested yesterday and charged with felony property damage for a rampage early Sunday at a Missouri State University parking lot. Investigators allege that Schaefer shattered car windows, ripped off side mirrors, dented hoods, and broke windshield wipers during the 3 AM spree. When confronted with surveillance footage and other evidence, Schaefer said that he could see the resemblance between the suspect and himself. At that point, Schaefer reportedly consented to a search of his iPhone. A subsequent review of the device revealed location data placing Schaefer "at or near the scene of the crime," as well as a "troubling dialogue exchange this defendant seems to have had with artificial intelligence software installed on his phone," prosecutors reported. The incriminating ChatGPT conversation can be found here.

Read more of this story at Slashdot.