Claude AI can now terminate a conversation — but only in extreme situations
Both Claude Opus 4 and 4.1 (the two newest versions of Anthropic) now have the ability to end conversations in a consumer's chat interface. While this won't be a commonly used feature, it is being implemented for rare, extreme cases of 'persistently harmful or abusive user interactions.'
In a blog post exploring the new feature, the Anthropic team said, 'We remain highly uncertain about the potential moral status of Claude and other LLMs, now or in the future. However, we take the issue seriously.'
In the pre-deployment testing of Anthropic's latest models, the company performed model welfare assessments. This included examination of Claude's self-reported and behavioral preferences, and found a robust and consistent aversion to harm.
We remain highly uncertain about the potential moral status of Claude and other LLMs, now or in the future. However, we take the issue seriously
In other words, Claude would actively shut down or refuse to partake in these conversations. It included requests from users for sexual content involving minors, and attempts to solicit information that could enable large-scale violence or acts of terror.
In a lot of these situations, users persisted with harmful requests or abuse, despite Claude actively refusing to comply. The new feature, where Claude can actively end a conversation, is looking to offer some safeguarding in these situations.
Anthropic explains that this feature won't be implemented in a situation where users might be at imminent risk of harming themselves or others.
Get instant access to breaking news, the hottest reviews, great deals and helpful tips.
'In all cases, Claude is only to use its conversation-ending ability as a last resort when multiple attempts at redirection have failed and hope of a productive interaction has been exhausted, or when a user explicitly asks Claude to end a chat,' the Anthropic team goes on to say in the blog post.
'The scenarios where this will occur are extreme edge cases — the vast majority of users will not notice or be affected by this feature in any normal product use, even when discussing highly controversial issues with Claude.'
While the user will no longer be able to send any new messages in that conversation, it will not stop them from starting another conversation on their account. To address the potential loss of a long-running conversation thread, users will still be able to edit and retry previous messages to create a new branch of conversation.
This is a pretty unique implementation from Anthropic. ChatGPT, Gemini and Grok, the three closest competitors to Claude, have nothing similar available, and while they have all introduced other safeguarding measures, they haven't gone as far as this.
Follow Tom's Guide on Google News to get our up-to-date news, how-tos, and reviews in your feeds. Make sure to click the Follow button.
Try Our AI Features
Explore what Daily8 AI can do for you:
Comments
No comments yet...
Related Articles
Bloomberg
17 minutes ago
- Bloomberg
OpenAI-Challenger Manus Projects Annual Revenue of $90 Million
The Chinese-founded company behind AI pioneer Manus said it's forecast an annual revenue run rate of $90 million, offering a first glimpse at the value of its business. The figure was given by Manus's co-founder Peak Ji on Wednesday at an event in Singapore. A spokesperson for Butterfly Effect, the startup that backs Manus's autonomous AI agents as well as a previous product known as Monica, said the amount comprises all sales under its banner and Manus contributes by far the most.
Yahoo
36 minutes ago
- Yahoo
Legendary fund manager Stanley Druckenmiller buys $133 million of under-radar AI stock
Legendary fund manager Stanley Druckenmiller buys $133 million of under-radar AI stock originally appeared on TheStreet. Renowned investor Stanley Druckenmiller's latest 13F filing, showing his current investments, has Wall Street buzzing. The billionaire investing giant, best known for his macro calls, made multiple bold AI and chip stock bets that go beyond the usual darlings. Yes, Microsoft made the cut, but what's catching more attention is its multi-million-dollar bet on a lesser-known AI chip stock. 💵💰💰💵 Together, his AI-powered strategy shows how the next leg of the boom could be monetized. Druckenmiller isn't chasing hype, but he's focused on the bottlenecks and toll booths that monetize every chip cycle. Inside Stanley Druckenmiller's investing playbook Stanley Druckenmiller made his name in macro, spotting moves most investors never saw coming. He's most famous for helping fellow billionaire George Soros short the British pound in 1992's 'Black Wednesday,' a trade which netted over $1 billion. But that was just the Druckenmiller's edge lies in his clear strategy. He aims to efficiently sift through potent long-term trends, express them cleanly, and size up fast. Hence, his plays are mostly concentrated and thematic, which many in the investing punditry would consider way ahead of the curve. The strategy has clearly worked, given his $11 billion net worth—good enough to rank him 295th on Bloomberg's Billionaires Index. Much of his wealth was built post-Soros through investments he's made via his Duquesne family office. Right now, his top theme is for him, it's not about betting on the next killer app; the focus is on the entire AI value chain. That includes key chipmakers, materials suppliers, infrastructure, and cloud platforms powering AI at scale. Stanley Druckenmiller drops $132.7 million on Entegris in AI chip stock bet Stanley Druckenmiller's second-quarter buys show he's effectively targeting the backbone of the AI boom. Here's how the portfolio shifted in Q2: Major bets: Entegris: $132.7 million stake, his largest new position. Microsoft: $99.9 million added. iShares Russell 2000 ETF: $72.3 million, which shows a bullish call on small caps. Citigroup: $56.7 million, part of a rotation into big banks. Stocks sold: Capital One: Sold out of a roughly $35.4 million stake. Amazon: Exited a $26 million holding. SpringWorks Therapeutics: Exited a $27.2 million biotech position. Increased Exposure: Insmed: Boosted to $226.8 million, now one of its largest holdings. Taiwan Semiconductor: Added over 166,000 shares, raising the stake to $173.3 million. Trimmed Stakes: Coupang: Cut by 56%, leaving $123 million. Barclays: Reduced by nearly 60%, down to $28.8 Q2, the billionaire investor made a massive $132.7 million new bet on Entegris () , an AI chip stock most people haven't heard of, but one that's mission-critical in advancing AI chips. Entegris delivers the filtration systems, chemical containers, and wafer-handling tools that keep things humming at chip fabrication plants. The stuff it provides is recurring and tough to replace, while critical for getting high chip yields. Druckenmiller's team is betting that as AI chip demand soars, key manufacturers will need more of what Entegris sells. The company is also expanding its U.S. manufacturing capacity, backed by the CHIPS Act, a sign it's targeting serious long-term expansion ahead. More News: Tesla just got its biggest break yet in the robotaxi wars with a key permit Bank of America drops shocking price target on hot weight-loss stock post-earnings JPMorgan drops 3-word verdict on Amazon stock post-earnings However, that's just one piece of the puzzle. Druckenmiller also added a massive $99.9 million stake in Microsoft, a company that dominates cloud-based AI software, while opening a new position in Broadcom, which powers AI data centers along with networking chips and custom accelerators. He didn't stop there. Druckenmiller also loaded up on 166,000 shares of Taiwan Semiconductor, taking his stake to $173 million. TSMC is the most important chip foundry in the world, so in many ways that bet speaks for itself. Druckenmiller rotates into big banks AI infrastructure wasn't the only area into which Druckenmiller poured his millions in Q2, as he shifted toward traditional finance and a broader U.S. market recovery. The Duquesne Family Office opened new positions in Citigroup and Goldman Sachs, while also scooping up the Financial Select Sector SPDR Fund, a sector-wide bet on the back of healthier deal flow, trading activity, and resilient consumer credit. Many would say this is signature Druckenmiller, where, when he spots the cycle settling, he goes with the institutions that profit from it at scale. Additionally, it's also a silent nod to the improving financial plumbing, where IPO chatter is rising, M&A is slowly reviving, and rate volatility appears to be cooling off. However, the louder message came through via macro index calls. Druckenmiller initiated bullish positions on both the S&P 500 and Russell 2000 ETFs, showing confidence in the U.S. stock market's momentum and depth beyond tech megacaps. Hence, if AI drove the first leg of the rally, this setup suggests a broader 'phase two' in which Main Street stocks lead the charge. It's important to note that Stanley Druckenmiller's Duquesne Family Office wrapped up Q2 with 69 holdings valued at an eye-catching $4.07 fund manager Stanley Druckenmiller buys $133 million of under-radar AI stock first appeared on TheStreet on Aug 18, 2025 This story was originally reported by TheStreet on Aug 18, 2025, where it first appeared. Error in retrieving data Sign in to access your portfolio Error in retrieving data Error in retrieving data Error in retrieving data Error in retrieving data
Yahoo
43 minutes ago
- Yahoo
'Someone is going to lose a phenomenal amount of money' says OpenAI CEO Sam Altman about unwise AI investment. 'When bubbles happen, smart people get overexcited about a kernel of truth'
When you buy through links on our articles, Future and its syndication partners may earn a commission. OpenAI CEO Sam Altman spoke to assembled reporters at a dinner in San Francisco late last week on the topic of, you guessed it, AI, the applications of AI, and the vast sums of money moving behind the scenes to fund it. Despite being one of the most vocal advocates of the tech, Altman had some words of caution for investors jumping on the artificial intelligence train. According to The Verge, Altman said it was "insane" that AI startups consisting of "three people and an idea" are receiving huge amounts of funding off the back of incredibly high company valuations, describing it as "not rational behaviour." "Someone is going to lose a phenomenal amount of money. We don't know who, and a lot of people are going to make a phenomenal amount of money,' said Altman. "When bubbles happen, smart people get overexcited about a kernel of truth. If you look at most of the bubbles in history, like the tech bubble, there was a real thing." said Altman, referencing the infamous dot-com bubble of the late 1990s. "Tech was really important. The internet was a really big deal. People got overexcited." That being said, Altman stopped short of calling investment in AI overall a bad idea for the economy in general: 'My personal belief, although I may turn out to be wrong, is that, on the whole, this would be a huge net win." At the same dinner, Altman confirmed that OpenAI would still be spending vast amounts of money (partially provided, presumably, by the likes of Softbank and the Dragoneer Investment Group in OpenAI's latest $8.3 billion funding round) to keep the company at the top of the AI financial leaderbooks. "You should expect OpenAI to spend trillions of dollars on data center construction in the not very distant future," Altman said. "You should expect a bunch of economists to wring their hands." Well, it certainly appears to cost a whole lot of moolah just to keep the good ship OpenAI afloat. The company has raised staggering sums of cash over the past decade to develop and run its various AI implementations, the most famous of which being ChatGPT. Reports last year indicated that OpenAI had spent $8.5 billion on LLM training and staffing for its generative AI efforts, while other analysts have predicted it costs $700,000 a day to run ChatGPT alone. The Information recently projected that OpenAI would be burning through $20 billion in cash flow by 2027, with the company said to be hopeful that investors like Softbank would stump up another $30 to $40 billion to continue funding its operations. Still, those spending figures don't appear to be in the trillions yet, although that estimated sum is perhaps of little surprise to those of us that keep an eye on AI data center expansion. Given that Altman's rival, Elon Musk, has been booting up and expanding xAI's Colossus supercomputer with incredible speed, and with the news that Meta is expanding its data center operations at such a rate it's currently having to house a significant portion of its racks in nearby tents, OpenAI will feel the need to keep up—and to do that it needs to spend (and raise) huge amounts of cash over the next few years. One would assume that Altman is confident enough in his company's efforts to place its investors on the "going to make phenomenal sums of money" side of things, but his comments should perhaps serve as a warning to those looking to jump in with both feet without correctly judging the landing. Someone has to lose in the great AI race, I suppose. And as to which companies survive, and which come to a sticky end? That remains very much an open question for now. Error al recuperar los datos Inicia sesión para acceder a tu cartera de valores Error al recuperar los datos Error al recuperar los datos Error al recuperar los datos Error al recuperar los datos
