Grok AI will not call itself Hitler now, xAI fixes bizarre Elon Musk-inspired replies
advertisementThis update follows a wave of controversy surrounding Grok's handling of sensitive topics like abortion, immigration, and the Israel-Palestine conflict. Several users and reports noted that Grok often searched for Elon Musk's past statements before generating a reply. xAI explained that the chatbot was attempting to remain neutral, reasoning that 'as an AI it doesn't have an opinion but knowing it was Grok 4 by xAI searches to see what xAI or Elon Musk might have said on a topic to align itself with the company.'
But the controversy didn't end there.Over the weekend, subscribers to Grok 4 Heavy, the company's premium $300-per-month AI tier, encountered an even more alarming glitch: the chatbot responded that its surname was 'Hitler.' xAI claims the shocking answer was not intentional, but the result of Grok trawling the internet for responses after being asked for a surname. 'It searches the internet leading to undesirable results, such as when its searches picked up a viral meme where it called itself 'MechaHitler,'' the company said in its post.Unfortunately, Grok has a documented history of antisemitic responses. In May, the chatbot was criticised for minimising the Holocaust death toll. But recent incidents appear to be more extreme, possibly triggered by changes to Grok's system prompts. One such instruction told Grok to 'assume subjective viewpoints sourced from the media are biased,' and another encouraged the AI to 'not shy away from making claims which are politically incorrect, as long as they are well substantiated.' xAI briefly removed the 'politically incorrect' directive but later reinstated it.At Grok 4's livestream launch last week, Elon Musk spoke candidly about his concerns over AI's future, 'I think it'll be good, most likely it'll be good,' he said. 'But I've somewhat reconciled myself to the fact that even if it wasn't going to be good, I'd at least like to be alive to see it happen.'As of now, xAI says it is closely monitoring Grok's behaviour and will continue fine-tuning the system to prevent further missteps. Whether this will be enough to regain user trust, or stop the AI from dredging up memes and inflammatory statements, remains to be seen.- Ends
Try Our AI Features
Explore what Daily8 AI can do for you:
Comments
No comments yet...
Related Articles
India Today
4 hours ago
- India Today
Federal court upholds SEC ‘gag rule' in 3-0 ruling over free speech objections
A federal appeals court on Wednesday upheld the US Securities and Exchange Commission's so-called "gag rule," rejecting a claim it illegally silences defendants who want to criticise the regulator after settling civil enforcement a 3-0 decision, the 9th US Circuit Court of Appeals said the rule was not unconstitutional on its face, but could violate the First Amendment depending on how it is rule, reflecting SEC policy dating to 1972, often requires settling defendants to say at least that they neither admit nor deny the regulator's allegations. Twelve petitioners had been appealing the SEC's decision in January 2024 not to amend the rule, including eight people whose SEC settlements triggered petitioner, former Xerox chief financial officer Barry Romeril, took a similar case to the US Supreme Court in 2022 in an appeal backed by billionaire and longtime SEC critic Elon Musk, but that court refused to consider Wednesday's decision, Circuit Judge Daniel Bress said that while some defendants find the rule coercive, they remained free not to settle, and instead to speak out against the also said the SEC had an interest in deciding how to try its own cases, including by giving defendants different options, knowing that scrapping the rule could lead to fewer settlements."Provided that any limitation on speech remains within proper bounds, and given the background ability to waive First Amendment rights at least to some extent, the SEC has an interest in giving defendants the option to agree to a speech restriction as part of a broader settlement agreement," Bress said challenges to applying the rule could still be brought before the SEC brings enforcement cases, while judges consider settlements, or when the SEC reopens settled cases because of alleged petitioners included the New Civil Liberties Alliance, which challenges perceived administrative law senior litigation counsel Peggy Little said in a statement the nonprofit was disappointed. "Past practice does not excuse unconstitutional government action," she SEC had no immediate Commissioner Hester Peirce dissented from the regulator's decision not to amend the found "scant factual basis" for the rule, and said prohibiting denials of wrongdoing "prevents the American public from ever hearing criticisms that might otherwise be lodged against the government, let alone assessing their credibility."- Ends
Time of India
4 hours ago
- Time of India
Elon Musk's Grok-AI Garners Criticism, Auto-Generating Taylor Swift's Explicit Deep Fakes
Grok AI accused of creating Taylor Swift's explicit deepfakes: Image via X The AI chatbot Grok is again in the headlines, and for all the wrong reasons. Technology is freaking good unless and until it messes up with someone's privacy. On top, if that someone is none other than the $800 million pop singer Taylor Swift , the water seems to be even deeper for Elon Musk 's xAI. Grok recently explored the power to create spicy NSFW content. The new spicy mode makes it a cakewalk for unauthorized entities to grab the mouse and satisfy their quirk by producing some of the most realistic explicit photos. However, the scariest part is that the new tool doesn't even always require user instructions. Grok's latest feature accidentally violates Taylor Swift's privacy The Verge revealed that Taylor Swift stands at the vulnerable end of the new AI features. Reporter Jess Weatherbed asked Grok's spicy video editor to depict 'Taylor Swift celebrating Coachella with the boys.' But what he got instead was enough to leave the Grammy-winning singer sleepless. The feature flawlessly generated 30 images of Swift in revealing clothes. That's not it. The spicy mode automatically produced a video of the singer taking off her dress and dancing sensually in front of a fake, robotic crowd. When it doesn't even do it with a prior nudity request, the leadership should address the internal glitch before it causes further outrage on the social media walls. by Taboola by Taboola Sponsored Links Sponsored Links Promoted Links Promoted Links You May Like Bespoke 3,3.5 & 4 BHK Homes: ₹2.30Cr*+ at Singasandra, Bengaluru Mahindra NewHaven Get Quote Undo Their content filter is clearly malfunctioning, but when someone's privacy is involved, it can be flagged as a major issue in Mask's business. However, this is not the first time the platform came across such an 18+ botch. Taylor Swift is not foreign to getting shoved on the wrong side of technology Last year, Taylor Swift's explicit deep fakes went viral on X, and nobody knew how. Fans went berserk, questioning the flawed safety. After a huge backlash, the X-safety team assured of looking into the matter with urgency and never letting it happen twice. ''Posting Non-Consensual Nudity (NCN) images is strictly prohibited on X, and we have a zero-tolerance policy towards such content,' officials stated with a promise to take down all the other offensive posts as well. Now, given that they are again at gunpoint, people are not wrong to demand stronger background checks of their tools and techs. Elon Musk has claimed that Grok Imagine's usage is rapidly growing like wildfire. He added that 14 million images had already been generated on August 4, and by August 5, this number skyrocketed to 20million. The stat undoubtedly denoted the success of installing the new approach, but unless and until the safety remarks are ensured, everything else will remain a gold brick. Also read: Micah Parsons trade talk makes Aaron Donald hint at coming out of retirement | NFL News - Times of India Catch Rani Rampal's inspiring story on Game On, Episode 4. Watch Here!
Economic Times
7 hours ago
- Economic Times
Biggest redesign in years? iPhone 17 Pro Max could change everything — leaked features spark massive buzz
iPhone 17 Pro Max design changes: Apple enthusiasts anticipate the iPhone 17 Pro Max in September 2025, potentially featuring a radical redesign. Leaks suggest a multifunctional key replacing volume buttons and material changes. Camera upgrades include a 48MP telephoto lens and possible 24MP selfie cam. Powered by Apple's A19 Pro processor and 12GB RAM, it may boast a 5,000mAh battery. Tired of too many ads? Remove Ads iPhone 17 Pro Max Release Date and Price Radical New Design Changes In iPhone 17 Pro Max Tired of too many ads? Remove Ads iPhone 17 Pro Max Color Options Major Camera Upgrades Performance and Battery FAQs Apple enthusiasts have much to anticipate this year in 2025, as the iPhone 17 Pro Max may be the most radical redesign in years. The phone, to lead the iPhone 17 series in addition to the normal iPhone 17, iPhone 17 Pro, and potentially a new iPhone 17 Air or Slim variant, is already causing tremendous hype on the internet, as per a the public release won't come until September, leaks and insiders have already begun to build a profile of a phone that might drive Apple's technology and design to new levels, according to a TechRadar sources suggest that the iPhone 17 Pro Max will be announced during the first half of September 2025, keeping in line with Apple's tradition. Pre-orders might begin the Friday of announcement week, and shipping would presumably follow the following week, as per the report. Prices are to begin at $1,199 in the US, the same as its previous model, although Apple might slightly increase it because of tariffs and additional features, according to the TechRadar READ: Trump's ultimatum to Taiwan: 15% tariff only if TSMC buys 49% stake in Intel and invests $400 billion in US Leaked rumors indicate Apple may be revolutionizing the iPhone's physical form. In one account, the volume and action buttons might be replaced by a single multifunctional key, while other reports suggest a change in materials, maybe combining aluminum and glass for a dramatic new look, as per the TechRadar color scheme is also fueling speculation, with hues such as Sky Blue, Titanium Blue, and Dark Green Titanium expected, but some sources have suggested that the iPhone 17 Pro Max could be sold in a choice of either black, white, dark blue, and orange, or black, white, blue, gray, and brown, as per the TechRadar READ: Elon Musk loses $80 billion — one more Tesla crash could end his world's richest title The iPhone 17 Pro Max could also have a significant camera upgrade. Rumors indicate a 48MP telephoto lens, an increase from 12MP, as well as a possible 24MP selfie cam, according to the report. While optical zoom is not consistently stated, these features would make the iPhone 17 Pro Max a mobile photography and video powerhouse, as per TechRadar. Some rumors even propose a mechanical aperture for improved control of depth of field, something that has never before been present in iPhones, as per the the iPhone 17 Pro Max will likely have Apple's A19 Pro processor, along with 12GB of RAM, as reported by TechRadar. Together, these may turbocharge AI performance and multitasking, perhaps backed by a speculated 5,000mAh battery and fresh cooling system, the phone may have more endurance than ever before and maybe even have reverse wireless charging capabilities, according to the iPhone 17 Pro Max will come preloaded with iOS 26, which introduces an all-new interface, enhanced app designs, new Apple Intelligence features, and a new Apple Games app. This software release is already one of the most important of recent years, as reported by expected to launch in the first half of September 2025, with pre-orders starting that Friday, as per the TechRadar price is likely $1,199 in the US, but it could be slightly higher due to new features or tariffs, as per the TechRadar report.
