Saying 'please' and 'thank you' to ChatGPT costs OpenAI millions, Sam Altman says
Being polite to your AI assistant could cost millions of dollars.
OpenAI CEO Sam Altman revealed that showing good manners to a ChatGPT model — such as saying 'please' and 'thank you' — adds up to millions of dollars in operational expenses.
Altman responded to a user on X (formerly Twitter) who asked how much the company has lost in electricity costs from people being polite to their models.
'Tens of millions of dollars well spent — you never know,' the CEO wrote.
Sounds like someone saw what operating system Hal did in '2001: A Space Odyssey' and is going to be nice to their AI assistant just in case. Experts have also found that being polite to a chatbot makes the AI more likely to respond to you in kind.
Judging from Altman's cheeky tone, that 'tens of millions' figure likely isn't a precise number. But any message to ChatGPT, no matter how trivial or inane, requires the AI to initiate a full response in real time, relying on high-powered computing systems and increasing the computational load — thereby using massive amounts of electricity.
AI models rely heavily on energy stored in global data centers — which already accounts for about 2% of the global electricity consumption. According to Goldman Sachs (GS), each ChatGPT-4 query uses about 10 times more electricity than a standard Google (GOOGL) search.
Data from the Washington Post suggests that if one out of every 10 working Americans uses GPT-4 once a week for a year (meaning 52 queries by 17 million people), the power needed would be comparable to the electricity consumed by every household in Washington, D.C. — for 20 days.
Rene Hass, CEO of semiconductor company ARM Holdings (ARM), recently warned that AI could account for a quarter of America's total power consumption by 2030. That figure currently is 4%.
Polite responses also add to OpenAI's water bill. AI uses water to cool the servers that generate the data. A study from the University of California, Riverside, said that using GPT-4 to generate 100 words consumes up to three bottles of water — and even a three-word response such as 'You are welcome' uses about 1.5 ounces of water.
For the latest news, Facebook, Twitter and Instagram.
Hashtags
Try Our AI Features
Explore what Daily8 AI can do for you:
Comments
No comments yet...
Related Articles
Business Insider
20 minutes ago
- Business Insider
Goldman Sachs Sticks to Their Hold Rating for Rumo SA (RUMOF)
Goldman Sachs analyst Bruno Amorim maintained a Hold rating on Rumo SA (RUMOF – Research Report) yesterday. The company's shares closed last Thursday at $4.57. Confident Investing Starts Here: Easily unpack a company's performance with TipRanks' new KPI Data for smart investment decisions Receive undervalued, market resilient stocks right to your inbox with TipRanks' Smart Value Newsletter According to TipRanks, Amorim is a 4-star analyst with an average return of 6.4% and a 58.65% success rate. Rumo SA has an analyst consensus of Hold. RUMOF market cap is currently $6.52B and has a P/E ratio of -30.85.
Business Insider
an hour ago
- Business Insider
An OpenAI exec says she was diagnosed with breast cancer and that ChatGPT has helped her navigate it
Kate Rouch, the chief marketing officer at OpenAI, shared on Friday that she was diagnosed with invasive breast cancer weeks after assuming the role, which she called her "dream job," in December. In a thread posted on X, Rouch said she was sharing her story to help other women, adding, "We can't control what happens to us--but we can choose how we face it. My biggest lesson: no one fights alone." Prior to joining OpenAI as the company's first CMO, Rouch was CMO at Coinbase and, before that, spent over a decade at Meta, including as vice president, global head of brand and product marketing. Rouch said she started treatment right around the Super Bowl in February, when OpenAI aired its first-ever ad, and that she has since gone through 13 rounds of chemotherapy while leading OpenAI's marketing team. She wrote that she is expected to make a full recovery. "It has been the hardest season of life — for me, for my husband, and for our two young children," Rouch said, adding she has been supported by OpenAI "at every step." "Silicon Valley can be brutal and transactional. And yet — I've never felt more held," she said, adding that "people showed up in incredible and unexpected ways." Rouch also said OpenAI's ChatGPT has helped her navigate her diagnosis and treatment, including by explaining cancer in a way that is age-appropriate for her kids, helping her manage the side effects of chemo, and creating custom meditations. "Experiencing our work as a patient has made OpenAI's mission feel more personal and important," she said. Rouch said she was sharing her story to encourage other women to "prioritize their health over the demands of families and jobs." "A routine exam saved my life. It could save yours, too," she said. Business Insider reached out to OpenAI for comment. Kevin Weil, the chief product officer at OpenAI, expressed support for Rouch in a reply to her thread. "We love you @kate_rouch!" he wrote. "Proud of you for telling your story and for being so full of fight."
Yahoo
an hour ago
- Yahoo
Trump-backed bill could block Tennessee's AI laws, threaten deepfake protections
NASHVILLE, Tenn. (WKRN) — President Donald Trump's 'big, beautiful bill' could change the way Tennessee regulates artificial intelligence. New language added to the legislation would block states from regulating AI or risk losing federal broadband funding. That means state laws like the Elvis Act, which protects artists from having their voices or images stolen by AI, would be put on pause. 'We are just hoping that we're not taking a couple steps backward since there's no federal regulation currently in place,' Kaley Bonett, a Nashville entertainment lawyer at Hall Booth Smith, said. MAY: House Republicans include a 10-year ban on US states regulating AI in 'big, beautiful' bill 'The songwriters — all — are entitled to revenue generated from the performances and distribution of that particular sound recording of their song, so if there's a deepfake imposing and competing with the authentic, then it's going to affect where the revenue share goes,' a Nashville managing partner and shareholder for Hall Booth Smith, Karl Braun, explained. 'It's going to be very difficult for, let's say, a songwriter to protect themselves legally from this evolving kind of internet Wild, wild west.' 'The big beautiful bill that the president is advancing does not replace those regulations with anything, which is almost unprecedented,' Plaintiffs Attorney at Spragens Law, David Kieley, said. Also at risk is the 'Preventing Deep Fakes Images Act,' which makes it a felony to post fake AI images designed to harm someone's reputation. Nashville-based meteorologist Bree Smith was targeted by AI imagery herself, explaining that someone used AI to take her face and put it on someone else's semi-naked body. '[This bill] just means that we are saying for 10 years we are going to let people potentially be victimized in this way, and that can't happen,' Smith said. 'It's not going to be any more okay for something like that to happen 10 years from now than it is for it to happen today.' 'Even looking prospectively at things we haven't really encountered yet, but this would stop the state from being able to regulate things like driverless cars in your town, driverless 18-wheelers on Tennessee highways,' Kieley added. 'We don't know really how this technology could affect public safety things like 9-1-1, dispatch.' A law signed just last week by Governor Lee banning AI-generated child sexual abuse material could also be blocked. ⏩ 'It would put child sex predators and big tech over the safety of our families,' Sen. Heidi Campbell (D-Nashville) said. 'This is really antithetical to the values that people across the aisle in Tennessee have been communicating with our legislation.' The president's bill passed in the House of Representatives. It's now in the Senate, where supporters hope to finalize it by July. Senator Marsha Blackburn has previously stated that Tennessee needs AI safeguards, saying, 'Until we pass something that is federally preemptive, we can't call for a moratorium on those things.' Copyright 2025 Nexstar Media, Inc. All rights reserved. This material may not be published, broadcast, rewritten, or redistributed.
