Reddit sues AI startup Anthropic for allegedly using data without permission
FILE PHOTO: Reddit app is seen on a smartphone in this illustration taken, July 13, 2021. REUTERS/Dado Ruvic/Illustration/File Photo
(Reuters) -Reddit sued the artificial intelligence startup Anthropic on Wednesday, accusing it of stealing data from the social media discussion website to train its AI models despite publicly assuring it wouldn't.
The complaint filed in San Francisco Superior Court is the latest battle over AI companies' alleged unauthorized use of third-party content. Anthropic's backers include Amazon.com and Google parent Alphabet.
"We disagree with Reddit's claims and will defend ourselves vigorously," an Anthropic spokesperson said.
According to the complaint, Anthropic has resisted entering a licensing agreement even as it trained its Claude chatbot on Reddit content, despite assuring last July it had blocked its bots from accessing Reddit's platform.
Reddit quoted Claude admitting it was "trained on at least some Reddit data" and did not know if that content was deleted.
It also said Anthropic's bots have accessed or tried to access Reddit content more than 100,000 times, undermining the company's allegedly styling itself as an AI "white knight" committed to trust and honesty.
"Anthropic refuses to respect Reddit's guardrails and enter into a license agreement," unlike Google and OpenAI, the complaint said.
By scraping content and using it for commercial purposes, Anthropic violated Reddit's user policy and "enriched itself to the tune of tens of billions of dollars," the complaint added.
In a statement, Reddit Chief Legal Officer Ben Lee said "we believe in an open internet," but AI companies need "clear limitations" on how they use content they scrape.
Reddit and Anthropic are based in San Francisco, about a 10-minute walk from each other.
The lawsuit seeks unspecified restitution and punitive damages, and an injunction prohibiting Anthropic from using Reddit content for commercial purposes.
Anthropic introduced its newest Claude models, Opus 4 and Sonnet 4, on May 22. Overall annualized revenue has reached $3 billion, two people familiar with the matter said last week.
The case is Reddit Inc v Anthropic PBC, California Superior Court, San Francisco County, No. CGC-25-524892.
(Reporting by Jonathan Stempel in New York; Additional reporting by Blake Brittain in Washington, D.C.; Editing by Bill Berkrot)
Try Our AI Features
Explore what Daily8 AI can do for you:
Comments
No comments yet...
Related Articles
The Star
8 hours ago
- The Star
Washington, Hanoi to fast-track trade negotiations
THE country and the United States have agreed to speed up trade talks, Hanoi said. Vietnam has the third-biggest trade surplus with the United States, after China and Mexico, and is anxious to address the imbalance to head off President Donald Trump's threatened 46% levy as part of his global tariff blitz. Vietnamese and US officials agreed at a meeting in Paris on Wednesday to 'focus maximum efforts to achieve the best results in the third technical negotiation round' scheduled for the first half of June, Vietnam's trade ministry said in a statement yesterday. Trade minister Nguyen Hong Dien affirmed Vietnam's 'determination and goodwill' to reach an agreement with the United States, while his American counterpart Jamieson Greer said a deal would be important for Washington as well as Hanoi, according to the statement. Meanwhile, a delegation from Vietnam's agriculture ministry visiting the United States signed memorandums of understanding (MOU) to buy raw materials for animal feed from Ohio worth US$600mil (RM2.5bil), the ministry said in a separate statement on its website yesterday. That comes after Vietnam announced five MOUs to buy products worth around US$800mil (RM3.4bil) from Iowa over three years, including soybean meal, corn, wheat, dried soybeans and dried distillers grains. The South-East Asian country was expected to sign deals with the United States to buy more than US$2bil (RM8.5bil) worth of agricultural products in total, the statement added. Vietnam has been seeking help to solve the tariff stand-off from several US tech and industry giants, including Lockheed Martin, SpaceX and Google. — AFP
The Star
8 hours ago
- The Star
Google's new AI-powered search has arrived. Proceed with caution
Recently, I asked Google to help me plan my daughter's birthday party by finding a park in Oakland, California, with picnic tables. The site generated a list of parks nearby, so I went to scout two of them out – only to find there were, in fact, no tables. 'I was just there,' I typed to Google. 'I didn't see wooden tables.' Google acknowledged the mistake and produced another list, which again included one of the parks with no tables. I repeated this experiment by asking Google to find an affordable carwash nearby. Google listed a service for US$25 (RM106), but when I arrived, a carwash cost US$65 (RM276). I also asked Google to find a grocery store where I could buy an exotic pepper paste. Its list included a nearby Whole Foods, which didn't carry the item. I wasn't doing traditional web searches on I was testing the company's new AI Mode, a tool that is similar to chatbots like ChatGPT and Google's Gemini, where users can type in questions to get answers. AI Mode, which is rolling out worldwide in the coming weeks, will soon appear as a tab next to your search results. The arrival of AI Mode underscores how new technology is redefining what it means to search for something online. For decades, a web search involved looking up keywords, like 'most reliable car brands,' to show a list of relevant websites. Now, with generative AI, the technology that powers chatbots by using complex language models to guess what words belong together, you can ask more specific questions or make complicated requests. That could include directing it to create a chart comparing the five most reliable 2025 sedans. Google, which has already been showing AI-generated summaries on its search pages for the past year, said AI Mode was a new frontier for search that would complement – but not yet replace – its traditional counterpart. 'We're really trying for AI Mode to be best at a new class of questions that are harder, more specific, and really the best for when you're going back and forth trying to get something done,' Robby Stein, a Google executive who oversees the search product team, said in an interview. The prominent placement of AI Mode on shows that AI is rapidly becoming unavoidable. Meta has added a chatbot, Meta AI, in Messenger, WhatsApp and Instagram, and Microsoft has integrated AI into its Bing search engine and its latest Surface computers. What's unique about AI Mode is that the technology stitches together data from Google's vast empire of Internet services to provide an answer to a query. When you type a question, it could pull data from search queries on location information on Google Maps and Google's shopping data on consumer products. To help assess whether AI is the future of search, I tested the new tool against traditional Google searches for a multitude of personal tasks over span of about a week, including shopping for a toddler car seat, preparing for a Memorial Day barbecue and understanding the plot twists of a popular video game. The results were mixed, with lots of hits but also lots of misses, so I encourage people to use AI Mode with caution. Here's how it went. AI Mode vs Google Search For each of my experiments, I opened AI Mode in one browser tab and with its traditional search bar in another. I typed the same query in each tab, then compared AI Mode's answers with Google's top list of search results. That helped determine whether AI Mode was more effective or I was better off clicking on search results to find the answers. Searching for Things and Places My earlier examples of picnic tables, a grocery item and a cheap carwash were similar in that they involved asking Google to find places or objects in the real world. Each of those queries prompted Google's AI to pull my location information and scan sources found on the web. – Google's AI Mode list included two parks with no picnic tables, but when I used to do the same search, its top three results included parks nearby that had tables. – Google's AI Mode suggested that the carwash I visited was US$25 based on one user review that mentioned this price. But a Google search showed several Yelp reviews of the business, where people reported a more accurate range of US$50 (RM212) to US$70 (RM297). – Google's AI Mode generated a list of grocery stores, including Whole Foods, that potentially sold the aji amarillo paste I needed to make Peruvian chicken for a Memorial Day barbecue. When I did a normal Google search for the paste nearby, the search engine took me to an Instacart listing confirming that one of the stores listed by AI Mode, Berkeley Bowl, carried the paste. Winner: Google search by a long shot. AI Mode's suggestions were sometimes accurate, but failing to check its answers could lead you down the wrong path and waste your time. Google said users of AI Mode could share feedback so it could quickly learn. 'It's early days, and these are technologies that are just starting to roll out now,' Stein said. 'As we learn about how to improve it, we'll improve it as quickly as possible.' Product research In another test, I asked Google's AI to help me research toddler car seats. This is where I saw the technology's potential to become very useful. Unlike a traditional web search, which would require me to read reviews of various car seat models and jot down a list including their pricing and features, AI Mode did all of this for me. I typed: 'I'm shopping for a convertible car seat. Create a table for me including popular models from Graco, Chicco and others and include pricing and main features.' Google immediately generated a handy chart to make comparing five car seats easy. There were some hiccups: Some information was missing from the table, and I noticed that the pricing was wrong for two of the seats. Still, it was simple for me to ask the AI to make corrections, and overall, picking a car seat with this bespoke chart sped up the process for me compared with the old-school method. I tested AI Mode to research other products like birthday gifts for a 1-year-old and the best electric toothbrush. The suggestions were useful. Winner: AI Mode. It's a nifty shopping tool, though it's still wise to do a Google search to double-check the prices. Pop culture After becoming a sleep-deprived father with the attention span of a goldfish, I got in the habit of reading summaries of movies and TV shows with convoluted plots. Recently, I finished a popular video game, Clair Obscur: Expedition 33 , which had a complex storyline. So I asked Google to summarise what had happened. Google gathered information from various video game blogs, Reddit posts and YouTube videos to piece together a cohesive summary of the game's plot and many twists. It was a satisfying recap. I tested AI Mode on other pieces of pop culture, like the Apple TV show Severance and HBO's The Last of Us , including how the latter show was different from the video game on which it's based. The tool generated similarly useful summaries. Winner: AI Mode. A traditional Google search will show you plenty of plot summaries of TV shows, games and movies on various sites. But sometimes you just want a quick and dirty bullet-pointed recap. Bottom line A traditional Google search is still best for the simple act of looking for things to do nearby, but AI Mode could prove to be a nifty tool for more tedious tasks like product research for online shopping – an instant chart comparing baby car seats is helpful, even if imperfect. Just always check the answers. As for whether this is the future of search, consumers will probably decide that over time. If most of you prefer to use AI Mode, it probably will gradually replace Google as we know it. I still prefer an old-school search, but my feelings could change one chart of baby gear at a time. – © 2025 The New York Times Company This article originally appeared in The New York Times
The Star
10 hours ago
- The Star
UK secures action from Amazon to tackle fake reviews
An Amazon logo is seen at its centre in Darlington, County Durham, Britain September 3, 2020. REUTERS/Lee Smith LONDON (Reuters) -Amazon has committed to do more to tackle fake reviews, including sanctioning British businesses that boost their ratings through bogus posts, the country's competition watchdog said on Friday. Following a four-year probe, the Competition and Markets Authority (CMA) said Amazon's undertakings also addressed its concerns about 'catalogue abuse' where sellers use the reviews of well-performing unrelated products to boost ratings and mislead customers. Companies that break the rules could be banned from selling on the e-commerce giant's platform altogether, while users who post fake reviews could be banned from posting reviews. Amazon also committed to robust processes to quickly detect and remove fake reviews and catalogue abuse, the CMA said. The regulator, which estimates that 90% of consumers use online reviews to inform purchases, began investigating Amazon and Google in 2021 over possible breaches of consumer protection law. It secured similar commitments from Google in January. "Thesenewcommitments matter and help set the standard," CMA boss Sarah Cardell said in a statement. The watchdog is conducting an initial sweep of review platforms following the release of new reviews guidance in April, seeking to identify platforms that may need to do more to comply with consumer law. Under its newly-granted powers, the CMA can independently decide whether consumer law has been infringed and take action in case of breaches, including issuing fines and ordering businesses to improve their practices. (Reporting by Muvija M, Editing by Paul Sandle)
