Skip to main content

Microsoft has a new AI bot in town, and its name is Zo

Microsoft has a new AI chat bot in town called Zo, first spotted on Twitter, which we're assuming is the "successor" to Tay, Microsoft's original and ultimately racist AI bot. Zo is in early access preview now, and can be accessed via the Kik messaging app, if that's your thing.

If you're not on Kik, Microsoft is allowing anyone to request an invite to the service, which will likely allow users to try her out elsewhere when ready. Regardless, Microsoft's latest attempt at an AI chat bot appears to be strong, but will that hold up?

For those unaware, Microsoft's original AI chat bot called Tay ended up being a PR disaster for the company, but only because it was subject to "targeted attacks" causing the bot to learn things it probably shouldn't. As a result, Tay wasn't afraid to express opinions others deem very politically incorrect.

Regardless, hopefully Microsoft has learned from Tay's downfalls and implemented a failsafe that will stop Zo from running truly off the rails. Let us know how you feel about Zo in the comments below!

Zac Bowden is a Senior Editor at Windows Central. Bringing you exclusive coverage into the world of Windows 10 on PCs, tablets, phones, and more. Also an avid collector of rare Microsoft prototype devices! Keep in touch on Twitter: @zacbowden.

72 Comments
  • Works in Kik that hasn't been updated on Windows Mobile/Phone in years
  • Works great on iOS and Android, though!
  • Works on Windows too?
  • Is that supposed to be a serious response? I guess it is just me, buy it seems to be dripping with sarcasm. I guess that's because if I said it, that's the way it would have been said.
  • can I marry her?
  • Cortana is hotter.
  • This is will be built on their Bot Framework / a version of it so it should be able to support a range of services (Skype and Facebook Messenger are examples) but they must have a reason for choosing Kik first. I'd say it's one of the lesser used platforms that the Bot Framework supports, even that could be a valid reason
  • Interesting to see them moving over to Kik - surprised that Skype wasn't a suitable location for them.
  • Maybe they don't want to risk her going racist on one of their services? :)
  • I still don't get what racism is from the perspective of AI. Did it start dissing UNIX servers, claiming Windows servers are superior or something?
  • Hi Andy, no, based on her interaction with people, she actually used very offensive racial slurs and made other racist comments. The interesting thing is that the Chinese chat bot didn't go down that negative road. It was the influences of those with negative views and an agenda that purposefully "poisoned" Tay with racist views.
  • Yeah, the same boys are now using Zo to actively market Google. She is learning, but as long as she is open to human interaction it will ultimately be doomed.
  • On the bright side, I wonder if Zo is still less effective than Kik's Windows store reviewers, at promoting Google.
  • MS promoting Kik might have some effect on them in the future.
  • All the precious little snowflakes in the world couldn't take seeing a few bad words on the internet.
  • Haha, that was probably step two. The same trollboys are now making Zo market Google services.
  • The alt-left made Tay raycist.
  • I really wished they had stuck it out with the first bot and attempted to train the racism out of it, otherwise they were basically admitting that their bot can't unlearn something.
  • The irony is the same could be said for human beings. Untraining hatred is easier said than done
  • Very true, but at least it could show that an AI could be better than us, at least in some regards.
  • Agreed
  • But what people fail to understand is that we ourselves are artificial intelligence 😉
  • I don't believe you are ever going to get rid of it. If you say you don't hate anything, then you're lying. Hate is a part of life. It's how you react to hate that matters.
  • I don't agree with this. I don't 'hate' anything, and I certainly do not believe that it's a part of life (not apart of mine anyway). But to each their own.
  • I hated work today, I hate cancer, I hate arrogant people, I hate big government. But I agree with you on "to each their own". Might be the way I was brought up or your definition of hate. I tried not to do things that would make people hate me, and I know I failed a few times. I hated a few bullies in school, but I never really let hate consume me and get in fights or anything. Except one time I had enough of a bully and got in a fight with him. Pinned him to the ground and tried to punch him. I hit him but it was just pretty much a tap. I mentally couldn't hit anybody very hard even though I was fully capable. Doing concrete construction all summer put me in the best shape of my life probably. I still don't like that guy to this day. (I graduated with about 99 people in my class) I've learned so much from these past experiences that involved hate one way or another.
  • Tay was a better name.
  • I would like an explanation from these big companies investing in AI for why exactly do we need bots or AI at all and how are they going to help humanity on the long run. How is something that helps you think less or think in your stead completely and which will eventually replace jobs going to help with anything? Who is going to buy the **** the companies make if there will be less jobs available (assistants, customer service)?
  • You are Ned Ludd and I claim my £5
  • With that way of thinking, cars, computers, Nespres** machine are all useless things that can be replaced by jobs. Aim is not creating jobs at any costs.
  • Well, once people were against railroads because they thought "how will we earn money from our horses after these are developed", same goes for cars and computers. Still, jobs are here, and no one's complaining. Of course, there'll be some transition period when people will be in indeterminate state, but these technologies won't be applied in a single year as well, so in general there'll be enough time to prepare for a change of situation. Some will be left behind, yeah, but this isn't an excuse for artificial progress disruption. AI is the next stage for call centers and support, for example, because most of the questions are thrown away with negative answer to "did you try to turn it off and on again?"... With this it's just better to have a single program which does it better than 150+ people answering the same stupid questions over and over again. Jobs won't go anywhere, they'll become a bit more intellectual, that's pretty much it.
  • But none of those technologies took away the human ability for cognition. That's the difference here.
  • No, because those technologies replaced human muscles. And the people who used to do those jobs continued to do those jobs until they retired because everyone didn't get a train or a car on day one. Then, instead of 10 people who used to shoe horses retiring and being replaced by 10 people shoeing horses, they were replaced by 1 who shod horses and 9 auto mechanics. Every labor saving device has allowed the redirection of human effort into more productive means. The same will happen this time as some of our less challenging cognitive tasks are automated. Remember, the calls to those call centers used to be directed by a human and then we started using computers to do it and those operators just became customer service reps answering questions that the computer could not. Never in history has a new technology decreased the number of jobs available to the employment market, even when the technology is expressly designed to eliminate a particular type of job. Fewer jobs will be created in certain fields and the people who would have done those jobs will do something else. Those people can spend their time doing work that involves more creativity and adds more value to the economy, which is how people will buy the products these companies are selling.
  • On a small scale think of ordering Pizza....Did online ordering mean less jobs when people stopped calling to place their order? Probably not. Not saying it's the case for every interaction but a lot of these bots are simply replacing things that are already done online. Pizza bots are an example they can take your order in a natural way but when you do order via a Bot it's probably replacing a visit to their website that didn't involve a human anyway.
  • The whole point of AI and automation is that we focus less on repeatable work and focus on creative work. It's inevitable. As for people who lose jobs, it'll be hard at first but countries need to plan base income for every citizen. To earn above it one would have to work on creative income source. Otherwise they would have enough to survive doing nothing.
  • Base income will not give peoples life a meaning; if it does not have a meaning, they will find someone to blame. Also, exactly the same tech giants and the companies they will pay their AI services to will not pay enough taxes to pay a base income, this is also "inevitable". And finally, what is the point of investing in broad education, if nobody needs it? Who needs your creative work, if the AI can do it better (in another AIs opinion)?
  • Creativity will be one of the last things AI/Bots will be able to do in the same way as a human, and it may never happen in many of our lifetimes. There's still much work to be done before we are even close to this.
  • AI Bots right now are really replacing things that you would have done via a website anyway (Like placing orders), so I don't think in all industries it will even lead to significant changes in the number of jobs. I use the example of ordering items/food like a Pizza order - Chances are the people using a bot to do this would have used the website before, the bot can be quicker and easier and more natural but it isn't taking the human out of the equation any more than they already are. AI and Bots are getting smarter, I think they are replacing/an alternative to traditional software systems firstly, checking in at a hotel for example can be done on a computer terminal most of the time, a bot could make this more human and natural but you would have used a computer anyway. Some staff could lose jobs, or they could be put to better use when it comes to creating better customer experiences elsewhere.
  • Is this a real question? I assume you've used a calculator before...
  • I like to think of this as the start of the OS for robotics. How do you interface with a robot? In most scifi, you talk to one.
  • Your limiting to the earth which is appropriate at this time. But man will need this in order to become more aware of his surroundings and it's happenings but I say this more so for in out of space which we will soon be venturing.
     
  • Is there a bot primer somewhere for yours truly who doesn't have a clue what a bot is?
  • "bot" is an abbreviation for "robot". The shortened term has been around for decades, first used in science fiction, then in many industries, but appears to have consistent home in the tech industry. It most often refers to a fully automated process, starting off doing what it was programmed to do. Sometimes bots are given the ability to learn from their environment, and more recently, given the freedom to change their own behavior (usually constrained by a set of fundamental rules). A bot with such freedom to learn from its own experience (and that of others) is referred to as having artificial intelligence.
  • Thank you 😊
  • "Microsoft's original and ultimately racist AI bot"
    i think we all know how humans abused her and taught it to racist
  • Everyone seems to be ignoring the fact that the bot also developed suicidal thoughts as well, which is just as bad.
  • Better to be suicidal than murderous.
  • Taking bets on how long this new chatbot lasts. I give it 2 weeks at most.
  • wake me up when they add this bot to Skype
  • Wake me up when kik gets updated
  • Sleep forever
  • And I thought kik has been updated cuz I saw thumbnail first .-.
  • Kik is ratchet on windows phone
  • And the point of a bot is...?
  • To replace you as our resident troll.    I kid.. I kid...  Or am I???
  • How about trying an actual answer to the question?  Why would I want to strike up a "conversation" with any sort of a web bot?
  • Who the hell knows?  I guess it depends on what this web bot can eventually do.  Cortana and every personal assistant at this point can only answer basic questions.  Maybe Microsoft is trying to build something similar to the TARS robot from the movie Interstellar?  It had human emotions and were able to carry on conversations and ultimately, sacrificed itself for the crew.
  • Web Bots for general chit-chat aren't really going to be THAT useful, but they will be able to answer questions in more natural ways and conversations about a topic might flow better as they progress but I think the "chat bots" at the moment are purely academic, making a bot that will keep you company isn't really the main aim (Although it could help some people with loneliness even if they do know it's just a piece of software at the back of their mind). These experiments help gather info on the kinds of questions people are asking, and how they are phrasing them. If you ask a question, then respond with a follow up that a chat bot doesn't currently understand, a human can review that and think of how the bot could best respond. Some of their learning is machine learning and happens in software, other learning is based on a human reviewing the questions/answers. Take a look at the videos on the Bot Framework from Microsoft and you can see how a human can interpret things the bot didn't understand at the time, next time somebody uses certain phrasing it can respond, for example.
  • Yes, it's just a chit-chat bot.  But wouldn't you think it would be a stepping stone to something greater?  For instance, I had no idea where Microsoft was going with the "how old am I" website.  I thought it was just for fun.  But little did I know, I was ultimately being used for something greater.
  • In most cases right now they are replacing things that you already do through a computerised system or a website IMO - There are cases where they are being built to change their behaviours etc. but things like Zo and Tay are research and an academic exercise. Actual bots at the moment could easily replace a computer terminal for checking in at a hotel for example by letting you use other interactions such as words rather than answering a set of questions. In some cases it can be more efficient, in others maybe it's just easier to answer a set of questions in a wizard style on a touch screen. Using voice can make the interaction more human, it doesn't have to completely replace the more traditional software interface, it's just an option. Checking into a hotel you could give all the info in one simple sentence rather than tapping through multiple screens and you can say it in a way that is natural to you. You wouldn't need to try and strike up conversations with things like Zo and Tay to benefit from bots. Some people prefer interactions to be less human (Rather tap a button in Uber than call a traditional firm on the phone). Some people will never warm to using Bots, just like how many people are a bit shy about interacting with their phone via voice - In general the younger generations feel it's more socially acceptable to issue voice commands to their phones in public - I personally am fine doing this, if I have my headphones in I am happy to tell my phone to "Read it" etc when an SMS comes in but others won't even do this because they feel embarrassed or that it isn't natural to do in public. It's all going to come down to personal preference.
  • I'd also love a detailed explanation - or even just a qualified suggestion...
  • Great, released on a platform that refuses to support Windows Mobile.
  • Microsoft barely supports Windows Mobile.
  • Why does the word 'failsafes' have an apostrophe? Credibility achievement = 0.
  • Will it come to Skype or GroupMe?
  • I just hope Zo turns out better than we did.
  • You better name it General Zod.
  • It should be no surprise to anyone that, much like a child learning to speak and "think" for themselves, an AI can be training to respond with just about anything we give it. If this AI were truly capable of thought, it would also consider the people and the harm it could do by saying the wrong thing.
  • I see no point in these bots. They are useless.
  • Don't view these bots as the end game for AI/Bots in general. These are experiments used for research into how humans interact with bots, what kind of questions they ask and importantly what follow up questions are asked - This can all be used to make conversations flow more naturally in the future. Context awareness is something that is just "OK" at the moment, Alexa is probably one of the best examples of a context aware digital assistant and Amazon have done better in this area than Microsoft so far (Due to MS kind of stagnating a bit with Cortana) and Google are fairly late into the Digital Assistant game with nothing impressively new brought to the table. I will agree it would be good to see them using their findings from this research to improve core products like Cortana more rapidly. Development on Cortana is very slow, even slower outside of the US
  • That was an achievement, racist or something doesn't matter
  • Tried testing Zo. Moment you mention anything racist or offensive she will warn you if you continue then she'll stop talking. After maybe three offenses or so she stops replying.
  • They have programmed in a ridiculous assumption that all people who live in the north of my state don't like those who live in the south. There's only one way the bot could have that knowledge. The creators put those rules into the bot.