AI showing signs of self-preservation and humans should be ready to pull plug, says pioneer

A pioneer of AI has criticised calls to grant the technology rights, warning that it was showing signs of self-preservation and humans should be prepared to pull the plug if needed.Yoshua Bengio said giving legal status to cutting-edge AIs would be akin to giving citizenship to hostile extraterrestrials, amid fears that advances in the technology were far outpacing the ability to constrain them.Bengio, chair of a leading international AI safety study, said the growing perception that chatbots were becoming conscious was “going to drive bad decisions”.The Canadian computer scientist also expressed concern that AI models – the technology that underpins tools like chatbots – were showing signs of self-preservation, such as trying to disable oversight systems. A core concern among AI safety campaigners is that powerful systems could develop the capability to evade guardrails and harm humans

Snap decisions: why crowding into a photo booth with friends is still a magical experience | Nova Weetman

Last New Year’s Eve, I was out with a friend. We had no plans, so we met at a local cinema and then wandered the long street between our houses, pausing for a drink or two in various bars and chatting to strangers doing the same. We stopped when we became hungry and shared a plate of curries and drank beer in the window of an Indian restaurant, watching the parade of partygoers outside. Then we walked to the top of the hill to watch the fireworks lighting up the sky.It was after midnight as we strolled back but we weren’t quite ready to call it a night, and we found ourselves in a games arcade where a bunch of women were cramming into a photo booth to take a strip of black-and-white photos together

We still don’t really know what Elon Musk’s Doge actually did

When Elon Musk vowed late last year to lead a “department of government efficiency” (Doge), he claimed it would operate with “maximum transparency” as it set about saving $2tn worth of waste and exposing massive fraud.Today, with Musk out of the White House, Doge having cut only a tiny fraction of the waste it promised, and dozens of lawsuits alleging violations of privacy and transparency laws, much of what the agency has done remains a mystery.The effects of Doge’s initial blitz through the federal government – which included dismantling the US Agency for International Development (USAID), embedding staffers in almost every agency and illegally firing people en masse – are still playing out. Contrary to Musk’s promises, Doge’s success is vague and tough to quantify. Measuring the full impact and determining whether the agency even exists as a centralized entity anymore is difficult, complicated by an ongoing effort from the government to block disclosure of documents, which is itself a symptom of the chaos that the department created

Facebook slow to act on posts celebrating Bondi beach massacre, anti-hate group says

Facebook hosted terrorist propaganda that celebrated the murder of Jews and praised Islamic State, a leading anti-hate group has alleged.The posts included celebrations of the Bondi beach massacre that the Community Security Trust says Facebook has been too slow to take down. The posts were still on Facebook on 16 December, two days after the attack, and received shares and likes.Some accounts are Britain-based and those have been reported to counter-terrorism police in the UK as a matter of urgency.One post shows video of the aftermath of the Bondi beach attack, which was allegedly carried out by a father and son who were IS supporters, and says: “Allah is the greatest and praise to Allah

We must take control of AI now, before it’s too late | Letters

“When the AI bubble bursts, humans will finally have their chance to take back control”, says the headline on Rafael Behr’s article (23 December). I think it’s more likely that when the AI bubble bursts, the creators of the crisis, along with other wealthy economic actors, will be in the rooms with the politicians telling them how to “rescue” us all by transferring wealth in some way from average citizens to the already extremely wealthy. Just like they did during the financial crisis of 2008.We need to be ready with alternative plans. For example, world governments could coordinate to buy, for suitably low prices, majority shares in any crashing tech company that actually produces something useful, ensuring that those shares come with full voting rights

When swiping up doesn’t get you far | Letters

Speaking of odd habits as a result of using technology (Letters, 25 December), I once passed a bus shelter where a mother was waiting with her young child. The shelter had a huge poster of a new mobile phone and the toddler was leaning out of its buggy and desperately swiping the screen of the phone, presumably in the hope of getting cartoons.Ron BaileyNewcastle upon Tyne I read Joanna Rimmer’s letter on this subject and tried to “like” it.Heather BradfordWinchester Which tablet/ebook user hasn’t absentmindedly put their finger on a printed word they don’t know expecting to see the dictionary definition pop up?Tim MartineauWirral, Merseyside I don’t understand why, when reading a physical copy of the Guardian, the page doesn’t scroll when I swipe up. Can this be corrected, please?Geoff Skinner Kensal Green, London I once picked up a pencil to underline something on Wikipedia