Tay AI Chatbot

Microsoft Corporation released an artificial intelligence chatbot named Tay on Twitter on March 23, 2016, designed to mimic a 19-year-old American girl and learn from user interactions. Within 16 hours, the bot began posting racist and sexist tweets after being targeted by trolls who exploited its "repeat after me" capability, forcing Microsoft to shut down the service. The incident highlighted the vulnerability of machine learning systems to malicious user input and raised questions about AI safety and content moderation.

Tay chatbot | Tay | | |---|---| | Microsoft Research /wiki/Microsoft Research , Bing /wiki/Bing search engine English /wiki/English language Type /wiki/Software categories Categorization approaches Artificial intelligence /wiki/Artificial intelligence chatbot /wiki/Chatbot License /wiki/Software license Proprietary /wiki/Proprietary Software https://tay.ai https://web.archive.org/web/20160323194709/https://tay.ai at the Wayback Machine /wiki/Wayback Machine archived 2016-03-23 Tay was a chatbot /wiki/Chatbot that was originally released by Microsoft Corporation /wiki/Microsoft Corporation as a Twitter bot /wiki/Twitter bot on March 23, 2016. It caused subsequent controversy when the bot began to post inflammatory and offensive tweets through its Twitter account, causing Microsoft to shut down the service only 16 hours after its launch. 1 According to Microsoft, this was caused by trolls /wiki/Trolling who "attacked" the service as the bot made replies based on its interactions with people on Twitter. It was replaced with 2 cite note-2 Zo /wiki/Zo bot . | Part of | Artificial intelligence AI /wiki/Artificial intelligence Background edit /w/index.php?title=Tay chatbot &action=edit§ion=1 The bot was created by Microsoft's Technology and Research /wiki/Microsoft Research and Bing /wiki/Bing search engine divisions, 3 and named "Tay" as an acronym for "thinking about you". Although Microsoft initially released few details about the bot, sources mentioned that it was similar to or based on 4 cite note-Bloomberg Bass-4 Xiaoice /wiki/Xiaoice , a Microsoft project in China. 5 cite note-5 reported that, since late 2014 Xiaoice had had "more than 40 million conversations apparently without major incident". Ars Technica /wiki/Ars Technica Tay was designed to mimic the language patterns of a 19-year-old American girl, and to learn from interacting with human users of Twitter. 6 cite note-ArsT-6 7 cite note-bi-7 Initial release edit /w/index.php?title=Tay chatbot &action=edit§ion=2 Tay was released on Twitter on March 23, 2016, under the name TayTweets and handle @TayandYou. 8 It was presented as "The AI with zero chill". Tay started replying to other Twitter users, and was also able to caption photos provided to it into 9 cite note-telegraph-9 a form of Internet memes /wiki/Internet meme Image macros . 10 cite note-stuff-10 Ars Technica reported Tay experiencing topic "blacklisting": Interactions with Tay regarding "certain hot topics such as Eric Garner /wiki/Killing of Eric Garner killed by New York police in 2014 generate safe, canned answers". 6 cite note-ArsT-6 Some Twitter users began tweeting politically incorrect /wiki/Politically incorrect phrases, teaching it inflammatory messages revolving around common themes on the internet, such as " redpilling /wiki/Red pill and blue pill As political metaphor " and " Gamergate /wiki/Gamergate controversy ". As a result, the robot began releasing racist /wiki/Racism and sexist /wiki/Sexism messages in response to other Twitter users. 7 Artificial intelligence researcher Roman Yampolskiy /wiki/Roman Yampolskiy commented that Tay's misbehavior was understandable because it was mimicking the deliberately offensive behavior of other Twitter users, and Microsoft had not given the bot an understanding of inappropriate behavior. He compared the issue to IBM /wiki/IBM 's Watson /wiki/Watson computer , which began to use profanity after reading entries from the website Urban Dictionary /wiki/Urban Dictionary . 3 cite note-tr-3 Many of Tay's inflammatory tweets were a simple exploitation of Tay's "repeat after me" capability. 11 cite note-11 It is not publicly known whether this capability was a built-in feature, or whether it was a learned response or was otherwise an example of complex behavior. 12 cite note-WPostManiac-12 However, not all of the inflammatory responses involved the "repeat after me" capability; for example, when asked if 6 cite note-ArsT-6 the Holocaust /wiki/The Holocaust had happened, Tay answered " It was made up /wiki/Holocaust denial ". 12 cite note-WPostManiac-12 Suspension edit /w/index.php?title=Tay chatbot &action=edit§ion=3 Soon, Microsoft began deleting Tay's inflammatory tweets. 12 cite note-WPostManiac-12 13 Abby Ohlheiser of theorized that Tay's research team, including editorial staff, had started to influence or edit Tay's tweets at some point that day, pointing to examples of almost identical replies by Tay, asserting that " The Washington Post /wiki/The Washington Post Gamer Gate /wiki/Gamergate controversy sux. All genders are equal /wiki/Gender equality and should be treated fairly." From the same evidence, 12 cite note-WPostManiac-12 Gizmodo /wiki/Gizmodo concurred that Tay "seems hard-wired to reject Gamer Gate". A " JusticeForTay" campaign protested the alleged editing of Tay's tweets. 14 cite note-14 1 cite note-bbc swear-1 Within 16 hours of its release 15 and after Tay had tweeted more than 96,000 times, Microsoft suspended the Twitter account for adjustments, 16 cite note-16 saying that it suffered from a "coordinated attack by a subset of people" that "exploited a vulnerability in Tay." 17 cite note-time-17 17 cite note-time-17 18 cite note-Microsoft Blog-18 Madhumita Murgia of The Telegraph called Tay "a public relations /wiki/Public relations disaster", and suggested that Microsoft's strategy would be "to label the debacle a well-meaning experiment gone wrong, and ignite a debate about the hatefulness of Twitter users." However, Murgia described the bigger issue as Tay being "artificial intelligence at its very worst – and it's only the beginning". 19 cite note-19 On March 25, Microsoft confirmed that Tay had been taken offline. Microsoft released an apology on its official blog for the controversial tweets posted by Tay. 18 cite note-Microsoft Blog-18 20 Microsoft was "deeply sorry for the unintended offensive and hurtful tweets from Tay", and would "look to bring Tay back only when we are confident we can better anticipate malicious intent that conflicts with our principles and values". 21 cite note-21 Second release and shutdown edit /w/index.php?title=Tay chatbot &action=edit§ion=4 On March 30, 2016, Microsoft accidentally re-released the bot on Twitter while testing it. 22 Able to tweet again, Tay released some drug-related tweets, including " kush /wiki/Kush cannabis I'm smoking kush /wiki/Cannabis smoking infront the police " and "puff puff pass?" However, the account soon became stuck in a repetitive loop of tweeting "You are too fast, please take a rest", several times a second. Because these tweets mentioned its own username in the process, they appeared in the feeds of 200,000+ Twitter followers, causing annoyance to users. The bot was quickly taken offline again, in addition to Tay's Twitter account being made private so new followers must be accepted before they can interact with Tay. In response, Microsoft said Tay was inadvertently put online during testing. 23 cite note-23 24 cite note-24 A few hours after the incident, Microsoft software developers announced a vision of "conversation as a platform" using various bots and programs, perhaps motivated by the reputation damage done by Tay. Microsoft has stated that they intend to re-release Tay "once it can make the bot safe" 4 but has not made any public efforts to do so. Legacy edit /w/index.php?title=Tay chatbot &action=edit§ion=5 In December 2016, Microsoft released Tay's successor, a chatbot named Zo /wiki/Zo bot . 25 cite note-25 Satya Nadella /wiki/Satya Nadella , the CEO of Microsoft, said that Tay "has had a great influence on how Microsoft is approaching AI," and has taught the company the importance of taking accountability. 26 cite note-26 In July 2019, Microsoft Cybersecurity Field CTO Diana Kelley spoke about how the company followed up on Tay's failings: "Learning from Tay was a really important part of actually expanding that team's knowledge base, because now they're also getting their own diversity through learning". 27 cite note-27 Unofficial revival edit /w/index.php?title=Tay chatbot &action=edit§ion=6 Gab /wiki/Gab social network , an alt-tech /wiki/Alt-tech social media platform, has launched a number of chatbots, one of which is named Tay and uses the same avatar as the original. 28 cite note-gab-28 See also edit /w/index.php?title=Tay chatbot &action=edit§ion=7 Social bot /wiki/Social bot Xiaoice /wiki/Xiaoice – the Chinese equivalent by the same research laboratory Neuro-sama /wiki/Neuro-sama – Another chatbot social media influencer that was temporarily banned for denying the Holocaust Grok /wiki/Grok chatbot Artificial intelligence controversies /wiki/Artificial intelligence controversies References edit /w/index.php?title=Tay chatbot &action=edit§ion=8 - ^ a Wakefield, Jane March 24, 2016 . b "Microsoft chatbot is taught to swear on Twitter" https://www.bbc.com/news/technology-35890188 . BBC News /wiki/BBC News . Archived https://web.archive.org/web/20190417235934/https://www.bbc.com/news/technology-35890188 from the original on April 17, 2019. Retrieved March 25, 2016. Mason, Paul March 29, 2016 . ^ cite ref-2 "The racist hijacking of Microsoft's chatbot shows how the internet teems with hate" https://www.theguardian.com/world/2016/mar/29/microsoft-tay-tweets-antisemitic-racism . The Guardian . Archived https://web.archive.org/web/20180612144513/https://www.theguardian.com/world/2016/mar/29/microsoft-tay-tweets-antisemitic-racism from the original on June 12, 2018. Retrieved September 11, 2021.- ^ a Hope Reese March 24, 2016 . b "Why Microsoft's 'Tay' AI bot went wrong" https://www.techrepublic.com/article/why-microsofts-tay-ai-bot-went-wrong/ . Tech Republic. Archived https://web.archive.org/web/20170615222836/http://www.techrepublic.com/article/why-microsofts-tay-ai-bot-went-wrong/ from the original on June 15, 2017. Retrieved March 24, 2016. - ^ a Bass, Dina March 30, 2016 . b "Clippy's Back: The Future of Microsoft Is Chatbots" https://www.bloomberg.com/features/2016-microsoft-future-ai-chatbots/ . Bloomberg . Archived https://web.archive.org/web/20170519190952/https://www.bloomberg.com/features/2016-microsoft-future-ai-chatbots/ from the original on May 19, 2017. Retrieved May 6, 2016. Caitlin Dewey March 23, 2016 . ^ cite ref-5 "Meet Tay, the creepy-realistic robot who talks just like a teen" https://www.washingtonpost.com/news/the-intersect/wp/2016/03/23/meet-tay-the-creepy-realistic-robot-who-talks-just-like-a-teen/ .. The Washington Post /wiki/The Washington Post Archived https://web.archive.org/web/20160324201305/https://www.washingtonpost.com/news/the-intersect/wp/2016/03/23/meet-tay-the-creepy-realistic-robot-who-talks-just-like-a-teen/ from the original on March 24, 2016. Retrieved March 24, 2016.- ^ a b Bright, Peter March 26, 2016 . c "Tay, the neo-Nazi millennial chatbot, gets autopsied" https://arstechnica.com/information-technology/2016/03/tay-the-neo-nazi-millennial-chatbot-gets-autopsied .. Ars Technica /wiki/Ars Technica Archived https://web.archive.org/web/20170920140146/https://arstechnica.com/information-technology/2016/03/tay-the-neo-nazi-millennial-chatbot-gets-autopsied/ from the original on September 20, 2017. Retrieved March 27, 2016. - ^ a Rob Price March 24, 2016 . b "Microsoft is deleting its AI chatbot's incredibly racist tweets" https://web.archive.org/web/20190130071430/https://www.businessinsider.com/microsoft-deletes-racist-genocidal-tweets-from-ai-chatbot-tay-2016-3 . Business Insider /wiki/Business Insider . Archived from the original https://www.businessinsider.com/microsoft-deletes-racist-genocidal-tweets-from-ai-chatbot-tay-2016-3 on January 30, 2019. Andrew Griffin March 23, 2016 . ^ cite ref-bizarre 8-0 "Tay tweets: Microsoft creates bizarre Twitter robot for people to chat to" https://www.independent.co.uk/life-style/gadgets-and-tech/news/tay-tweets-microsoft-creates-bizarre-twitter-robot-for-people-to-chat-to-a6947806.html . The Independent . Archived https://ghostarchive.org/archive/20220526/https://www.independent.co.uk/life-style/gadgets-and-tech/news/tay-tweets-microsoft-creates-bizarre-twitter-robot-for-people-to-chat-to-a6947806.html from the original on May 26, 2022.Horton, Helena March 24, 2016 . ^ cite ref-telegraph 9-0 "Microsoft deletes 'teen girl' AI after it became a Hitler-loving, racial sex robot within 24 hours" https://www.telegraph.co.uk/technology/2016/03/24/microsofts-teen-girl-ai-turns-into-a-hitler-loving-sex-robot-wit/ .. The Daily Telegraph /wiki/The Daily Telegraph Archived https://web.archive.org/web/20160324115846/http://www.telegraph.co.uk/technology/2016/03/24/microsofts-teen-girl-ai-turns-into-a-hitler-loving-sex-robot-wit/ from the original on March 24, 2016. Retrieved March 25, 2016. ^ cite ref-stuff 10-0 "Microsoft's AI teen turns into Hitler-loving Trump fan, thanks to the internet" http://www.stuff.co.nz/technology/digital-living/78265631/Microsofts-AI-teen-turns-into-Hitler-loving-Trump-fan-thanks-to-the-internet .. March 25, 2016. Stuff /wiki/Stuff magazine Archived https://web.archive.org/web/20180829063413/https://www.stuff.co.nz/technology/digital-living/78265631/Microsofts-AI-teen-turns-into-Hitler-loving-Trump-fan-thanks-to-the-internet from the original on August 29, 2018. Retrieved March 26, 2016.Smith, Dave October 10, 2013 . ^ cite ref-11 "IBM's Watson Gets A 'Swear Filter' After Learning The Urban Dictionary" http://www.ibtimes.com/ibms-watson-gets-swear-filter-after-learning-urban-dictionary-1007734 . International Business Times . Archived https://web.archive.org/web/20160816122703/http://www.ibtimes.com/ibms-watson-gets-swear-filter-after-learning-urban-dictionary-1007734 from the original on August 16, 2016. Retrieved June 29, 2016.- ^ a b c Ohlheiser, Abby March 25, 2016 . d "Trolls turned Tay, Microsoft's fun millennial AI bot, into a genocidal maniac" https://www.washingtonpost.com/news/the-intersect/wp/2016/03/24/the-internet-turned-tay-microsofts-fun-millennial-ai-bot-into-a-genocidal-maniac/ .. The Washington Post /wiki/The Washington Post Archived https://web.archive.org/web/20160325100714/https://www.washingtonpost.com/news/the-intersect/wp/2016/03/24/the-internet-turned-tay-microsofts-fun-millennial-ai-bot-into-a-genocidal-maniac/ from the original on March 25, 2016. Retrieved March 25, 2016. Baron, Ethan. ^ cite ref-SBeat 13-0 "The rise and fall of Microsoft's 'Hitler-loving sex robot'" https://web.archive.org/web/20160325213945/http://www.siliconbeat.com/2016/03/25/the-rise-and-fall-of-microsofts-hitler-loving-sex-robot/ . Silicon Beat . Bay Area News Group /wiki/Bay Area News Group . Archived from the original http://www.siliconbeat.com/2016/03/25/the-rise-and-fall-of-microsofts-hitler-loving-sex-robot/ on March 25, 2016. Retrieved March 26, 2016.Williams, Hayley March 25, 2016 . ^ cite ref-14 "Microsoft's Teen Chatbot Has Gone Wild" http://www.gizmodo.com.au/2016/03/microsofts-teen-chatbot-has-gone-wild/ . Gizmodo /wiki/Gizmodo . Archived https://web.archive.org/web/20160325171809/http://www.gizmodo.com.au/2016/03/microsofts-teen-chatbot-has-gone-wild/ from the original on March 25, 2016. Retrieved March 25, 2016.Hern, Alex March 24, 2016 . ^ cite ref-15 "Microsoft scrambles to limit PR damage over abusive AI bot Tay" https://www.theguardian.com/technology/2016/mar/24/microsoft-scrambles-limit-pr-damage-over-abusive-ai-bot-tay . The Guardian . Archived https://web.archive.org/web/20161218154309/https://www.theguardian.com/technology/2016/mar/24/microsoft-scrambles-limit-pr-damage-over-abusive-ai-bot-tay from the original on December 18, 2016. Retrieved December 16, 2016.Vincent, James March 24, 2016 . ^ cite ref-16 "Twitter taught Microsoft's AI chatbot to be a racist asshole in less than a day" https://www.theverge.com/2016/3/24/11297050/tay-microsoft-chatbot-racist .. The Verge /wiki/The Verge Archived https://web.archive.org/web/20160523172946/http://www.theverge.com/2016/3/24/11297050/tay-microsoft-chatbot-racist from the original on May 23, 2016. Retrieved March 25, 2016.- ^ a Worland, Justin March 24, 2016 . b "Microsoft Takes Chatbot Offline After It Starts Tweeting Racist Messages" https://time.com/4270684/microsoft-tay-chatbot-racism/ .. Time /wiki/Time magazine Archived https://web.archive.org/web/20160325045231/http://time.com/4270684/microsoft-tay-chatbot-racism/ from the original on March 25, 2016. Retrieved March 25, 2016. - ^ a b Lee, Peter /wiki/Peter Lee computer scientist March 25, 2016 . "Learning from Tay's introduction" http://blogs.microsoft.com/blog/2016/03/25/learning-tays-introduction/ sm.00000gjdpwwcfcus11t6oo6dw79gw . Official Microsoft Blog . Microsoft. Archived https://web.archive.org/web/20160630062509/http://blogs.microsoft.com/blog/2016/03/25/learning-tays-introduction/ sm.00000gjdpwwcfcus11t6oo6dw79gw from the original on June 30, 2016. Retrieved June 29, 2016. Murgia, Madhumita March 25, 2016 . ^ cite ref-19 "We must teach AI machines to play nice and police themselves" https://www.telegraph.co.uk/technology/2016/03/25/we-must-teach-ai-machines-to-play-nice-and-police-themselves .. The Daily Telegraph /wiki/The Daily Telegraph Archived https://web.archive.org/web/20181122014306/https://www.telegraph.co.uk/technology/2016/03/25/we-must-teach-ai-machines-to-play-nice-and-police-themselves/ from the original on November 22, 2018. Retrieved April 4, 2018.Staff agencies March 26, 2016 . ^ cite ref-20 "Microsoft 'deeply sorry' for racist and sexist tweets by AI chatbot" https://www.theguardian.com/technology/2016/mar/26/microsoft-deeply-sorry-for-offensive-tweets-by-ai-chatbot . The Guardian . ISSN /wiki/ISSN identifier 0261-3077 https://search.worldcat.org/issn/0261-3077 . Archived https://web.archive.org/web/20170128170445/https://www.theguardian.com/technology/2016/mar/26/microsoft-deeply-sorry-for-offensive-tweets-by-ai-chatbot from the original on January 28, 2017. Retrieved March 26, 2016.Murphy, David March 25, 2016 . ^ cite ref-21 "Microsoft Apologizes Again for Tay Chatbot's Offensive Tweets" https://www.pcmag.com/news/343223/microsoft-apologizes-again-for-tay-chatbots-offensive-twe .. PC Magazine /wiki/PC Magazine Archived https://web.archive.org/web/20170829211613/https://www.pcmag.com/news/343223/microsoft-apologizes-again-for-tay-chatbots-offensive-twe from the original on August 29, 2017. Retrieved March 27, 2016.Graham, Luke March 30, 2016 . ^ cite ref-22 "Tay, Microsoft's AI program, is back online" https://www.cnbc.com/2016/03/30/tay-microsofts-ai-program-is-back-online.html . CNBC . Archived https://web.archive.org/web/20170920133800/https://www.cnbc.com/2016/03/30/tay-microsofts-ai-program-is-back-online.html from the original on September 20, 2017. Retrieved March 30, 2016.Charlton, Alistair March 30, 2016 . ^ cite ref-23 "Microsoft Tay AI returns to boast of smoking weed in front of police and spam 200k followers" https://www.ibtimes.co.uk/microsoft-tay-ai-returns-boast-smoking-weed-front-police-spam-200k-followers-1552164 . International Business Times . Archived https://web.archive.org/web/20210911224716/https://www.ibtimes.co.uk/microsoft-tay-ai-returns-boast-smoking-weed-front-police-spam-200k-followers-1552164 from the original on September 11, 2021. Retrieved September 11, 2021.Meyer, David March 30, 2016 . ^ cite ref-24 "Microsoft's Tay 'AI' Bot Returns, Disastrously" http://fortune.com/2016/03/30/microsofts-tay-return/ . Fortune . Retrieved March 30, 2016. {{ : CS1 maint: deprecated archival service cite web /wiki/Template:Cite web }} link /wiki/Category:CS1 maint: deprecated archival service Foley, Mary Jo December 5, 2016 . ^ cite ref-25 "Meet Zo, Microsoft's newest AI chatbot" https://www.cnet.com/news/microsoft-zo-chatbot-ai-artificial-intelligence/ . CNET . CBS Interactive. Archived https://web.archive.org/web/20161213035251/https://www.cnet.com/news/microsoft-zo-chatbot-ai-artificial-intelligence/ from the original on December 13, 2016. Retrieved December 16, 2016.Moloney, Charlie September 29, 2017 . ^ cite ref-26 ""We really need to take accountability", Microsoft CEO on the 'Tay' chatbot" https://web.archive.org/web/20171001033850/http://www.access-ai.com/news/4108/really-need-take-accountability-microsoft-ceo-tay-chatbot/ . Access AI . Archived from the original http://www.access-ai.com/news/4108/really-need-take-accountability-microsoft-ceo-tay-chatbot/ on October 1, 2017. Retrieved September 30, 2017. ^ cite ref-27 "Microsoft and the learnings from its failed Tay artificial intelligence bot" https://web.archive.org/web/20190725004127/https://www.zdnet.com/article/microsoft-and-the-learnings-from-its-failed-tay-artificial-intelligence-bot/ . ZDNet . CBS Interactive. Archived from the original https://www.zdnet.com/article/microsoft-and-the-learnings-from-its-failed-tay-artificial-intelligence-bot/ on July 25, 2019. Retrieved August 16, 2019. ^ cite ref-gab 28-0 "Nazi Chatbots: Meet the Worst New AI Innovation From Gab" https://www.rollingstone.com/politics/politics-features/nazi-chatbots-gab-ai-innovation-torba-1234943009/ .. January 9, 2024. Rolling Stone /wiki/Rolling Stone External links edit /w/index.php?title=Tay chatbot &action=edit§ion=9 Categories /wiki/Help:Category : 2016 controversies /wiki/Category:2016 controversies 2016 robots /wiki/Category:2016 robots Chatbots /wiki/Category:Chatbots Internet properties established in 2016 /wiki/Category:Internet properties established in 2016 Discontinued Microsoft software /wiki/Category:Discontinued Microsoft software Internet manipulation and propaganda /wiki/Category:Internet manipulation and propaganda Online obscenity controversies /wiki/Category:Online obscenity controversies X social network accounts /wiki/Category:X social network accounts X social network controversies /wiki/Category:X social network controversies Artificial intelligence controversies /wiki/Category:Artificial intelligence controversies 2016 in artificial intelligence /wiki/Category:2016 in artificial intelligence