The Supreme Court Battle for Section 230 Has Begunupstart writes:The future of recommendation algorithms could be at stake:The first shots have been fired in a Supreme Court showdown over web platforms, terrorism, and Section 230 of the Communications Decency Act. Today, the Supreme Court will hear oral arguments in Gonzales v. Google — one of two lawsuits that are likely to shape the future of the internet.Gonzalez v. Google and Twitter v. Taamneh are a pair of lawsuits blaming platforms for facilitating Islamic State attacks. The court's final ruling on these cases will determine web services' liability for hosting illegal activity, particularly if they promote it with algorithmic recommendations.The Supreme Court took up both cases in October: one at the request of a family that's suing Google and the other as a preemptive defense filed by Twitter. They're two of the latest in a long string of suits alleging that websites are legally responsible for failing to remove terrorist propaganda. The vast majority of these suits have failed, often thanks to Section 230, which shields companies from liability for hosting illegal content. But the two petitions respond to a more mixed 2021 opinion from the Ninth Circuit Court of Appeals, which threw out two terrorism-related suits but allowed a third to proceed.Gonzalez v. Google claims Google knowingly hosted Islamic State propaganda that allegedly led to a 2015 attack in Paris, thus providing material support to an illegal terrorist group. But while the case is nominally about terrorist content, its core question is whether amplifying an illegal post makes companies responsible for it. In addition to simply not banning Islamic State videos, the plaintiffs — the estate of a woman who died in the attack — say that YouTube recommended these videos automatically to others, spreading them across the platform.Google has asserted that it's protected by Section 230, but the plaintiffs argue that the law's boundaries are undecided. "[Section 230] does not contain specific language regarding recommendations, and does not provide a distinct legal standard governing recommendations," they said in yesterday's legal filing. They're asking the Supreme Court to find that some recommendation systems are a kind of direct publication — as well as some pieces of metadata, including hyperlinks generated for an uploaded video and notifications alerting people to that video. By extension, they hope that could make services liable for promoting it.Read more of this story at SoylentNews.
fliptop writes:Under the fluorescent lights of a fifth grade classroom in Lexington, Kentucky, Donnie Piercey instructed his 23 students to try and outwit the "robot" that was churning out writing assignments:
upstart writes:The change comes after early beta testers of the chatbot found that it could go off the rails and discuss violence, declare love, and insist that it was right when it was wrong:
canopic jug writes:OpenBSD developer, Florian Obser, has written a detailed post on privilege drop, privilege separation, and restricted-service operating mode in OpenBSD. The BSD-derived operating project, OpenBSD, has been at the forefront of mitigation techniques, for decades now. Florian discusses what OpenBSD has now, and how it got there and provides examples.
An Anonymous Coward writes:Australian science people and a bunch of teachers have discovered three new spider species in a remote alpine region. Given Australia has so many spiders, and indeed other deadly creatures, it is a wonder they look for more. One of the new spiders discovered is a 'jumping spider' that 'looks like a "liquorice-all-sort" and is a "ferocious hunter" that camouflages as a tree branch'. In typical Australian style, another looks like an eight eyed puppy. No, really.
Motor Trend and probably many other sources report on the recall of all Teslas with FSD, https://www.motortrend.com/news/tesla-admits-full-self-driving-crashes-recalls-360000-cars/
quietus writes:It is now legal to hack into any company or government institution in Belgium, but only under certain circumstances.That's the result of a new law on whistleblower protection that came into effect on February 15. Under the law, any citizen with the Belgian nationality is allowed to breach networks of Belgian legal entities without any previous notification or consent, provided he/she describes and reports the breach to the Centre for Cybersecurity Belgium within 72 hours, and does not request a reward for it.
Mykl writes:An article over at The Register describes how Bing's new Ai powered Chat service (currently in a limited Beta test) lied, denied, and claimed a hoax when presented with evidence that it was susceptible to Prompt Injection attacks. A user named "mirobin" posted a comment to Reddit describing a conversation he had with the bot: