Need to get something done online, but don’t have the time or energy? There could soon be a bot to do the hard work for you, as David Crookes explains
What’s a bot?
Bots are software applications that have been programmed to perform tasks automatically over the internet or to pretend to be a person. The word is short for ‘web robot’ and the idea is that they perform simple, time-consuming and repetitive functions that would otherwise require some form of human involvement. They’ve been around for nearly as long as the internet itself but they’re set to become an even more important part of the way we use the web in the coming months and years.
What kinds of bots are there?
Bots have various uses. Google, for example, uses them to crawl and catalogue the web, which is how it’s able to deliver relevant search results, while criminals use them to assault the internet with automated attacks and look for unprotected computers. Other bots help people win online auctions by bidding for them at the last second, and give online gamers an advantage by performing repetitive tasks in games. However, it’s chatbots – pieces of software designed to simulate conversation – that are currently grabbing the headlines.
Why have chatbots been grabbing the headlines?
There have been plenty of developments in the field lately. For a start, we’ve seen the emergence of bot platforms and bot stores for the messaging apps Telegram (telegram.org) and Kik (www.kik.com), which let you chat to bots about a whole range of topics including sport, fashion and the weather. The big news has come from Microsoft and Facebook, however. Both companies are looking at ways to incorporate chatbots into their messaging platforms and are inviting third-party companies to get involved. Microsoft CEO Satya Nadella has gone so far as to say “bots are the new apps”.
That’s a bold claim. But why the emphasis on chatbots?
There’s a big push towards getting us to talk to machines rather than clicking and tapping our way around menus and buttons. Chatbots have simple user interfaces that let us access services by the most natural means we have: our own language. These bots can gather data, learn our habits and even work out what we want to do before we realise it. Currently, you can use bots to order food, book flights, receive tech support or find out if the trousers you like are available in a particular shop. But there’s a sense that we’re just at the beginning of the bot era.
Why would Microsoft describe bots as the new apps?
It would appear that the app market is reaching a saturation point. Figures released in January by Forrester Research (bit.ly/forrester396) showed that the average smartphone owner in the US spends 88 per cent of their time using just five apps: Facebook, YouTube, Instagram, Gmail and Facebook Messenger. With people reluctant to cram more apps on their devices, companies are looking for alternative ways to attract attention and gain business. Social-media and messaging apps show little sign of declining in popularity, so a lot of thought is going into how people can use simple messages to communicate with online services. Hence the rise of the bots.
Won’t we need apps to use bots?
This is where things become interesting. Rather than having a string of bots built into thousands of websites and apps, we can anticipate a rising number of platforms that bring various bots together – a kind of onestop bot hub. This certainly seems to be what Microsoft and Facebook want to do. Indeed, at its F8 conference, Facebook’s head of messaging, David Marcus, launched the Messenger Platform (messengerplatform.fb.com), which will allow any company to produce its own bots. Although the bots could be embedded on a website, Facebook would rather users communicated with them directly through its Messenger app.
What bots are active on Messenger right now?
For a while, Facebook Messenger users in the US have been able to book, track and pay for an Uber taxi without needing to download an extra app or leave a conversation. Now additional developers are on board, and have been creating everything from a weather bot called Poncho to a CNN bot that delivers news when you want it. Other companies that are getting involved include Burger King, eBay, Expedia, Staples and US online clothing store Spring, each of which offers conversational ways of getting things done.
The idea is that anyone who wants to carry out a task could end up heading straight to Messenger. For example, you could ask a hotel bot: “Can I book a room tonight?” and receive a message stating the current availability and giving you the chance to reply with a confirmation. There would be no need to visit the hotel’s website or app, and you could go straight from that to messaging another service or even a friend to let them know where you’re staying.
How is Microsoft using bots?
At the Microsoft Build 2016 conference in March, the company unveiled a more intelligent version of its digital personal assistant, Cortana. Intended to work with ‘conversational platforms’ such as Skype, the new Cortana can eavesdrop on your text chats and automatically pick up on any tasks that may need to be performed. Because Cortana has access to third-party bots, it could pop up at the appropriate moment and make a suggestion. You could then have a side conversation with the bot using natural language, and get things done quickly.
How would this be different to Messenger’s bots
Messenger’s bots will reach out to the web, whereas Microsoft’s will also be able to work with the files on your PC. Let’s say you’re chatting with a colleague about going on a business trip. Cortana would identify your plans, place a provisional entry in your PC’s diary and then broker third-party hotel and transport service bots to help you make the necessary travel arrangements. During this whole process, no human other than yourself would be involved.
Isn’t this all rather creepy?
Potentially. The ultimate aim is that, in conversing with many of these bots, you could be fooled into thinking you were speaking to a human sales rep rather than artificial intelligence. Having a bot interrupt your private conversation could quickly become annoying, especially if the messages seem like adverts or spam. Bots could also lead to job losses, especially in call centres (although they would eliminate call-waiting times). None of this should stop the development of bots but we may have to start asking some deeper ethical questions further down the line.
Haven’t we seen bots before?
Yes. The SmarterChild chatbot, which was active on AOL Instant Messenger and MSN Messenger around 2001, attracted 30 million people. Some have also pointed out parallels with the dreaded Clippy Assistant in Microsoft Office, which terrorised PC and Mac users from 1997 to 2003. Chatbots are certainly nothing new, but the current resurgence in interest – thanks to the popularity of social media and messaging apps – could truly be a game changer this time.
WHEN GOOD BOTS TURN BAD
Not all bots have proved successful. In March this year, in a bid to refine its voice-recognition software, Microsoft introduced an Artificial Intelligence chatbot on Twitter called Tay, who was supposed to speak like a teenage girl. The idea was that Tay would learn from the conversations people had with her, but she soon fell prey to trolls and troublemakers. As some users used “fruity” language and tried to lead Tay astray, she went off the rails, expressing her support for Hitler, blaming George W Bush for 9/11 and denying the Holocaust.
Microsoft responded by protecting Tay’s tweets to all but confirmed followers, and the experience hasn’t put the company off experimenting with bots. Shortly afterwards, it released CaptionBot (captionbot.ai), which lets you upload photos to find out what’s in them, and Murphy (projectmurphy.net), which lets you ask a series of ‘What if’ questions to see the result.