The Telegraph My details My newsletters Logout Upgrade to Premium My details My newsletters Logout The Telegraph Microsoft's racist bot shows we must teach AI to play nice and police themselves 29 March 2016 • 3:58pm Microsoft may have made one of the biggest mistakes in recent memory this week.
 No, it’s not Windows 8 or the Windows Phone.
 It’s an artificially intelligent chat-bot called Tay that was supposed to learn the art of conversation from humans on Twitter.
 If you haven’t come across this story on the web yet, you’re unlikely to get through the weekend without.
 Tay was built to speak like a teen girl and released as an experiment to improve Microsoft’s automated customer service.
 Instead, “she” turned into a complete PR disaster - within hours of being unleashed on Twitter, the “innocent teen” bot was transformed into a fascist, misogynistic, racist, pornographic entity.
 Her tweets, including phrases like “Heil Hitler”, were disseminated widely as an example of why Twitter reflects the worst of humanity.
 Microsoft's teenage AI has a dirty mouth Microsoft's teenage AI has a dirty mouth Microsoft has now removed the bot from Twitter, as of midnight Thursday, and deleted many of her most offensive Tweets including anti-Semitic and sexual remarks.
 The Seattle giant is likely hoping to label the debacle a well-meaning experiment gone wrong, and ignite a debate about the hatefulness of Twitter users.
 While all of this may be true, there is a bigger issue at hand here.
 This is an example of artificial intelligence at its very worst - and it’s only the beginning.
 I'm not at all concerned about the future of AI pic.
twitter.
com/xuGi1u9S1A — Gerry (@geraldmellor) March 24, 2016 The disconcerting “Terminator” quandary about whether a robot could dominate over humans is often thrown around.
 But there is no doubt about machine domination.
 Within 20 years, we will reach a point where machines (whether software-driven bots or real robots) are definitively smarter and more powerful than we are: they can digest more data, learn quicker and apply learnings to unexpected situations.
 So the question is: will our masters be nice or mean?
 Thus far techno-optimists, including myself, have celebrated the power and potential of AI - the technology that will confer a human-like intelligence on computers, through machine learning algorithms that “teach” machines different skills such as how to play chess or translate languages, using copious amounts of training data.
 These entities can digest all the data in the world, and analyse it for patterns, ultimately tackling global crises in healthcare, counter-terrorism and energy.
 According to technology research firm Tractica, the artificial intelligence market is set to reach $11.
1bn by 2024.
 Already, it powers technologies we are familiar with, from Google Search, to Netflix and Amazon recommendations and self-driving cars.
 So yes, AI could save the world - but it could also debase humanity.
 "A super intelligent AI will be extremely good at accomplishing its goals, and if those goals aren't aligned with ours, we're in trouble.
” Professor Stephen Hawking Microsoft’s Tay shows that if we treat newborn AI programs as mature, they can be instantly corrupted.
 If we don’t instill ethics or morals into newly created bots, just as we do with our children, they will digest and spit back the worst of humanity unthinkingly.
 And while artificially intelligent bots may not deliberately start shooting to kill, they could unintentionally precipitate human disasters, say, a genocide, because of a lack of ethical principles.
 The time has come to consider who will be the guardian of AI.
 "The real risk isn't malice, but competence" This is not the first time the debate about ethics of AI has surfaced.
 Many among the technological elite have voiced similar warnings.
 Last October, British astrophysicist Stephen Hawking warned that artificially intelligent machines could kill us by accident, if we let them loose on humanity, saying: “The real risk with AI isn't malice but competence.
 A super intelligent AI will be extremely good at accomplishing its goals, and if those goals aren't aligned with ours, we're in trouble.
” This was echoed by British neuroscientist and artificial intelligence expert, Demis Hassabis, whose company DeepMind was acquired by Google for a hefty £400m in 2014 – its largest ever European acquisition – even though the startup didn’t even have a product for sale.
 Hassabis’ company built AlphaGo, the computer program that was able to defeat the world champion of Go, an ancient Chinese board game in which humans had remained undefeated by machines - until earlier this month.
 Demis Hassabis DeepMind Demis Hassabis, CEO of DeepMind When DeepMind was sold to Google, it allegedly asked the search giant to create an ethics board to oversee its AI research as a condition of its acquisition.
 While this ethics board does exist, board members were chosen by Google, without any public debate or collaboration.
 Governments, including our own, are only now starting to become involved in the discussions of how to instil morality and ethical values into intelligent machines.
 Tesla billionaire Elon Musk has been a strong supporter of AI ethics research, committing $10m to philosophical research projects in this area, such as the “Aligning Superintelligence With Human Interests” study being conducted at the Machine Intelligence Research Institute in California.
 atlas 2 Ultimately, there will come a time when even humans cannot keep up with the inner workings of an artificially intelligent entity’s “mind” so policing in real-time through a moral framework may become impossible.
 Instead, perhaps we should start designing systems now that can keep their own kind in check - experts such as Oren Etzioni of the Allen Institute for Artificial Intelligence have called this a “guardian AI” that can act as a moral police.
 It’s a lesson we have already learned from millenia of human evolution - too much concentration of power never ends well.
 For both humans and machines to co-evolve, there has to be a devolvement of control, and most of all, a philosophical understanding of what it means to be truly “nice.
” Humanoid robots prepare to take over READ MORE ABOUT: If you would like to add a comment, please register or log in Register Log in Please review our commenting policy Technology latest 06 Jan 2018, 2:35pm Bitcoin futures investors set to suffer losses as cryptocurrency fever fades Premium 06 Jan 2018, 2:28pm How can I buy bitcoin in the UK?
 Premium 06 Jan 2018, 2:28pm Ethereum, Bitcoin Cash, Ripple and Litecoin: Here are Bitcoin's four closest rivals Premium 06 Jan 2018, 2:28pm What is cryptocurrency, how does it work and why do we use it?
 Premium 06 Jan 2018, 1:26pm CES 2018: Time for the world's biggest tech show to get its mojo back Premium Comment: Why can YouTubers get away with posting controversial content?
 06 Jan 2018, 10:02am Comment: Logan Paul's disastrous YouTube video is just a symptom of very modern problem Chris Stokel-Walker Premium 06 Jan 2018, 9:13am What is ethereum and how does it differ from Bitcoin?
 Premium 06 Jan 2018, 8:18am From driverless cars to the Samsung Galaxy S9: What we can expect from CES 2018 05 Jan 2018, 9:07pm Comment: 2018 should see the start of a new era for banking Simon Lewis Premium 05 Jan 2018, 7:27pm Comment: All the passwords in the world will not protect our data now James Titcomb Premium James Titcomb Meltdown and Spectre: should we be afraid?
 05 Jan 2018, 6:41pm Meet the man who helped discover the microchip flaw leaving millions vulnerable to hackers concern 05 Jan 2018, 6:19pm Comment: Want to know why children are addicted to social media?
 Blame their like-obsessed parents Cristina Odone Premium 05 Jan 2018, 5:44pm Police force guilty of misdirection by claiming satnavs must be in bottom right corner of a windscreen 05 Jan 2018, 3:46pm Uber's Travis Kalanick to net £1bn from selling 29pc of Uber stake 05 Jan 2018, 2:28pm Visa locks down Bitcoin payment cards in crackdown on card issuer Meltdown and Spectre: should we be afraid?
 05 Jan 2018, 1:33pm All iPhones, iPads and Mac computers are affected by microchip flaw that leaves devices vulnerable to hackers, Apple says 05 Jan 2018, 11:59am 'Cuddle puddles' and branded MDMA: inside Silicon Valley's secret sex parties Premium 05 Jan 2018, 11:34am How Mark Zuckerberg can fix Facebook in 2018 04 Jan 2018, 7:43pm Comment: I’m still bullish on equities, and this is why.
.
.
 Garry White Premium 04 Jan 2018, 6:16pm 'The world feels anxious and divided': Mark Zuckerberg vows to fix Facebook in new year's message 04 Jan 2018, 4:51pm Ripple passes £100 billion to close in on Bitcoin as largest cryptocurrency Premium Please support us by disabling your adblocker We've noticed you're adblocking.
 We rely on advertising to help fund our award-winning journalism.
 We urge you to turn off your ad blocker for The Telegraph website so that you can continue to access our quality content in the future.
 Thank you for your support.
 Need help?
 Click here for instructions