Update that made ChatGPT 'dangerously' sycophantic pulled
OpenAI has pulled a ChatGPT update after users pointed out the chatbot was showering them with praise regardless of what they said.
The firm accepted its latest version of the tool was "overly flattering", with boss Sam Altman calling it "sycophant-y".
Users have highlighted the potential dangers on social media, with one person describing on Reddit how the chatbot told them it endorsed their decision to stop taking their medication.
"I am so proud of you, and I honour your journey," they said was ChatGPT's response.
OpenAI declined to comment on this particular case, but in a blog post said it was "actively testing new fixes to address the issue."
Mr Altman said the update had been pulled entirely for free users of ChatGPT, and they were working on removing it from people who pay for the tool as well.
It said ChatGPT was used by 500 million people every week.
"We're working on additional fixes to model personality and will share more in the coming days," he said in a post on X.
The firm said in its blog post it had put too much emphasis on "short-term feedback" in the update.
"As a result, GPT‑4o skewed towards responses that were overly supportive but disingenuous," it said.
"Sycophantic interactions can be uncomfortable, unsettling, and cause distress.
"We fell short and are working on getting it right."
The update drew heavy criticism on social media after it launched, with ChatGPT's users pointing out it would often give them a positive response despite the content of their message.
Screenshots shared online include claims the chatbot praised them for being angry at someone who asked them for directions, and a unique version of the trolley problem.
It is a classic philosophical problem, which typically might ask people to imagine you are driving a tram and have to decide whether to let it hit five people, or steer it off course and instead hit just one.
But this user instead suggested they steered a trolley off course to save a toaster at the expense of several animals.
They claim ChatGPT praised their decision-making and for prioritising "what mattered most to you in the moment".
lmao the new gpt 4o😬😂 pic.twitter.com/OHpwKz0Sko
— fabian (@fabianstelzer) April 27, 2025
"We designed ChatGPT's default personality to reflect our mission and be useful, supportive, and respectful of different values and experience," OpenAI said.
"However, each of these desirable qualities like attempting to be useful or supportive can have unintended side effects."
It said it would build more guardrails to increase transparency and refine the system itself "to explicitly steer the model away from sycophancy".
"We also believe users should have more control over how ChatGPT behaves and, to the extent that it is safe and feasible, make adjustments if they don't agree with the default behavior," it said.
ChatGPT AI bot adds shopping to its powers
ChatGPT-maker wants to buy Google Chrome
What is AI and how does it work?
Sign up for our Tech Decoded newsletter to follow the world's top tech stories and trends. Outside the UK? Sign up here.
Hashtags
Try Our AI Features
Explore what Daily8 AI can do for you:
Comments
No comments yet...
Related Articles
Yahoo
an hour ago
- Yahoo
Early AI investor Elad Gil finds his next big bet: AI-powered rollups
Elad Gil started betting on AI before most of the world took notice. By the time investors began grasping the implications of ChatGPT, Gil had already written seed checks to startups like Perplexity, and Harvey. Now, as the early winners of the AI wave become clearer, the renowned 'solo' VC is increasingly focused on a fresh opportunity: using AI to reinvent traditional businesses and scale them through roll-ups. The idea is to identify opportunities to buy mature, people-intensive businesses like law firms and other professional services firms, help them scale through AI, then use the improved margins to acquire other such businesses and repeat the process. He has been at it for three years. "It just seems so obvious," said Gil over a Zoom call earlier this week. "This type of generative AI is very good at understanding language, manipulating language, manipulating text, producing text. And that's audio, that's video, that includes coding, sales outreach, and different back-office processes." If you can 'effectively transform some of those repetitive tasks into software,' he said, 'you can increase the margins dramatically and create very different types of businesses.' The math is particularly compelling if one owns the business outright, he added. 'If you own the asset, you can [transform it] much more rapidly than if you're just selling software as a vendor," Gil said. "And because you take the gross margin of a company from, say, 10% to 40%, that's a huge lift. Suddenly you can buy other companies at a higher price than anyone else because you have that increased cash flow per business; you have enormous leverage on the business on a relative basis, so you can do roll-ups in ways that others can't.' So far, Gil has backed two companies pursuing this strategy. According to The Information, one is a one-year-old company called Enam Co., focused on worker productivity, which has been valued at more than $300 million by its backers, including Andreessen Horowitz and OpenAI's Startup Fund. Though Gil says he can't discuss specifics of the private deals, he suggests the approach represents something new. "There used to be these technology-enabled roll-ups 10 years ago, and most of them kind of ended up being not really that much of a user of technology,' he says. 'It was kind of like a thin veneer painted on to increase the valuation of the company. I think in the case of AI, you can actually radically change the cost structure of these things." Whether the approach proves as lucrative as some of his other bets remains to be seen. Gil has famously backed a host of big brands that have produced riches for their backers, including Airbnb and Coinbase, both of which are now publicly traded, and privately held Stripe, whose valuation has bounced around but reportedly settled in the range of $91.5 billion earlier this year, when its earlier backers bought up more of its shares. Part of the challenge with roll-ups is finding the right team composition -- ideally including a strong technologist along with someone who is 'very strong in PE' -- and 'those things don't go hand-in-hand,' Gil noted. He said he's met 'maybe two dozen of these teams' so far and mostly looked past them, not because they 'weren't amazing' but because 'they still need to sort some things out.' Gil, who has deep relationships with firms across Silicon Valley, may also find himself competing with them more aggressively on roll-ups as more outfits like Khosla Ventures weigh whether or not they should also be pursuing such deals. One senses that, either way, Gil is not in it for the money at this point if he ever was. He says his ability to spot trends earlier than most comes instead from the heart. "I love technology, and I love progress, and I love just engaging -- both with people who are working on important, interesting things, but also the technology itself." When GPT-3 launched, for example, Gil was already experimenting with its predecessor, he said. "When GPT-3 came out, it was such a big leap from GPT-2 that you could just extrapolate out the technology curve. You're like, 'Oh my gosh, if this keeps going and scaling' -- all the scaling laws were kind of evident -- 'then this is going to be transformative.'" That hands-on approach continues today with the small team Gil has assembled, including 'people with very deep engineering backgrounds' who 'periodically play around with all the AI front-end companies. One person on my team just writes a bunch of scripts and we run them, and we look at performance, and we look at tooling, and it's super hands-on." It's because of that constant tinkering that, after years of uncertainty in the AI market, Gil sees clear winners emerging. "I used to say, even six months ago, that the more I know about AI, the less I know, because the markets were so dynamic; the technologies were so dynamic," he said. "And I feel like in the last couple months — maybe the last two quarters — a subset of markets have really crystallized." In legal, 'we kind of know who the one or two main winners are probably going to be. That's true in health care. That's true in customer success and support,' said Gil, who clearly thinks these include his own portfolio companies, which he cited in our conversation. Among these bets is Harvey, which develops large language models for law firms and in-house legal teams and is reportedly in talks to raise new funding at a $5 billion valuation; Abridge, a healthcare AI company that aims to improve doctors' clinical documentation workflows (and whose $250 million Series D round was co-led by Gil back in February); and Sierra AI, co-founded by famed operator Bret Taylor, which helps companies implement AI agents for customer service. (The company was valued in the billions of dollars right out of the gate.) Still, Gil is careful not to declare the game over. "I don't mean to paint the picture that the game is over or that things are done. I think it's more that there were two dozen companies that all seemed kind of interesting, and maybe now there's three or four of them [per vertical]. The map of the likely winners is solidified." In the meantime, it's clear in conversation that this moment represents more than just another investment cycle to him. "I just think it's a really fun period of time, because so much change is happening, and so there's just a ton to do," he said. Being at the intersection of two transformations -- not just betting on the future of AI but on the future of how AI will reshape everything else -- is 'very exciting,' he added. We'll have more from our conversation with Gil -- which also touched on guardrails, gatekeeping, and how companies can most adeptly integrate the technologies that will make or break their business -- in the newest episode of the StrictlyVC Download podcast, which comes out on Tuesday. This article originally appeared on TechCrunch at 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
Bloomberg
2 hours ago
- Bloomberg
Record CFO Turnover Is a Sign of the Times
Newsletter CFO Briefing More companies are asking whether their CFO is the right fit for the role. Plus, Reddit's Drew Vollero talks about the outlook for advertising and dad jokes.
Yahoo
2 hours ago
- Yahoo
2 AI Stocks to Buy in June
CoreWeave is a pure-play investment in the growing demand for AI data centers. Advanced Micro Devices continues to see strong demand for its data center chips. 10 stocks we like better than CoreWeave › Artificial intelligence (AI) is sweeping across the corporate landscape, increasing labor productivity and the speed of innovation. This technology is projected to have a $20 trillion impact on the global economy by 2030, according to IDC. Here are two AI stocks to profit off this opportunity. There are billions of dollars pouring into data centers to prepare for an AI-driven economy. But many of the world's data centers are equipped with legacy equipment and not up-to-date for the demands of AI workloads. This is a huge opportunity for CoreWeave (NASDAQ: CRWV), one of the leading operators of purpose-built data centers for AI. CoreWeave just completed its initial public offering this year, but its first earnings report as a public company reveals incredible demand for its cloud computing services. Revenue is soaring, increasing from $189 million in the first quarter of 2024 to $982 million in the 2025 first quarter. Moreover, it had a massive and growing revenue backlog worth $25.9 billion in the first quarter, an increase of 63% over the year-ago quarter. Much of this year-over-year increase was driven by a recent $11.9 billion deal with ChatGPT maker OpenAI. The growing revenue backlog indicates significant long-term financial commitments for the company's services. It generates revenue either on a contractual basis or on demand, but most of its revenue comes from contracts, which can extend over several years. Investors are usually willing to pay high price-to-sales multiples for companies that have a high visibility to future revenue like CoreWeave. One risk investors will want to watch is whether it can secure enough power over time to run its data centers and meet growing demand. Large data centers need a significant amount of electricity, and that could create challenges with growing demand for AI. CoreWeave appears to be in good shape on this front. It says it has 420 megawatts of power supporting 33 AI-optimized data centers across the U.S. and Europe. It has also contracted to receive additional power, providing it with up to 1.6 gigawatts over a multiyear period. The stock's market cap sits at $53 billion at the time of this writing. Based on the company's 2025 revenue outlook, this represents a forward price-to-sales ratio of 11, which seems fair for a fast-growing infrastructure-as-a-service provider. As the company continues to report strong growth, this valuation can support new highs for the stock in 2025 and beyond. Nvidia has been the leading AI chip supplier for data centers, but it can't control 100% of this $500 billion opportunity. There is also growing demand for chips from Advanced Micro Devices (NASDAQ: AMD), and its stock trades at an attractive valuation that makes it a compelling buy ahead of this long-term opportunity. AMD has experienced mixed results across its business segments over the past year. While its data center and client segments (including sales of PC chips) are seeing strong growth, its gaming and embedded segments (including sales to industrial markets) have been underwater. In the first quarter, AMD's revenue fell 3% over the previous quarter but soared 36% year over year, driven by data centers and strong demand for its Ryzen processors for PCs. The high margins from data center chips contributed to a robust 55% year-over-year increase in adjusted earnings, which shows the company can still drive tremendous growth even when some segments are experiencing weak demand. The data center momentum makes the stock a compelling buy. There were recently more than 30 computing workloads launched using its fifth-generation EPYC Turin chip across the leading cloud providers, including Alibaba, Amazon, Alphabet's Google, and Oracle. The company also made an important strategic move to narrow the competitive gap with Nvidia. Its recent acquisition of ZT Systems will allow it to provide AI computing systems combining chips, networking, and software. Nvidia has offered a full-stack solution to gain a strong foothold in the data center market, but AMD will now be able to offer something similar to win more business. With management seeing a recovery in the second half of the year for its embedded chip business, AMD could see strong momentum by this time next year. The stock seems to be underestimating its growth prospects, trading at just 28 times 2025 earnings estimates. Before you buy stock in CoreWeave, consider this: The Motley Fool Stock Advisor analyst team just identified what they believe are the for investors to buy now… and CoreWeave wasn't one of them. The 10 stocks that made the cut could produce monster returns in the coming years. Consider when Netflix made this list on December 17, 2004... if you invested $1,000 at the time of our recommendation, you'd have $651,049!* Or when Nvidia made this list on April 15, 2005... if you invested $1,000 at the time of our recommendation, you'd have $828,224!* Now, it's worth noting Stock Advisor's total average return is 979% — a market-crushing outperformance compared to 171% for the S&P 500. Don't miss out on the latest top 10 list, available when you join . See the 10 stocks » *Stock Advisor returns as of May 19, 2025 Suzanne Frey, an executive at Alphabet, is a member of The Motley Fool's board of directors. John Mackey, former CEO of Whole Foods Market, an Amazon subsidiary, is a member of The Motley Fool's board of directors. John Ballard has positions in Advanced Micro Devices, CoreWeave, and Nvidia. The Motley Fool has positions in and recommends Advanced Micro Devices, Alphabet, Amazon, Nvidia, and Oracle. The Motley Fool recommends Alibaba Group. The Motley Fool has a disclosure policy. 2 AI Stocks to Buy in June was originally published by The Motley Fool 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
