The botification of news

“If I ask my AI personal assistant, ‘What’s the most important news story of today?’ how will it determine what that is for me, and where will it go looking to deliver me the story?”

Bots have been around a long time on the web. Without them, Google wouldn’t function — its automated web spider bots are what enables it to constantly index and rank websites. The CAPTCHA forms you often have to fill in for online services are there to make sure you’re not a bot.

trushar-barotHowever, it will be the use of bots in news that will be a major development in 2016. Already this year, bots are starting to have an impact — mostly in chat apps. Kik, the Canadian chat app that has a large user base in the U.S. and is growing in other parts of the west, has developed a chatbot platform that enables users to ping it with key terms and get automated responses.

NBC launched a chatbot inside the platform earlier in the year, where it delivers news stories based on the key terms that a user sends to the bot as a message. At this stage, it’s still a pretty rudimentary experience, but other chat apps are making similar moves.

Telegram launched its bots in June this year. As a completely open source platform, it offers the most interesting applications so far, with several news organisations — including the BBC — experimenting with new ways of distributing and generating news for its users.

Other chat apps including Line and WeChat have bots which will also be developed further in 2016.

But it’s not just chatbots. Next year, we’ll start seeing the emergence of fully fledged AI personal assistants. Facebook is already well on the way to developing its version called M, which will live inside Messenger. M will be able to book reservations for you, buy products online, and retrieve information you need. Apple has Siri, Google has Google Now, Microsoft has Cortana — but its most interesting AI bot is Xiaoice, a text-based bot for Chinese users.

It uses machine learning to scour the Chinese Internet and learns how normal people talk, constantly improving it’s algorithm. It can remember personal anecdotes users share. For example, Xiaoice may ask if you’re feeling better about a breakup you mentioned in a previous conversation with her (yes, the chatbot has a gender).

Why is this of interest to news organizations? Well, we spend a lot of time with SEO and social media optimisation. But how will we operate in a world where a personal AI bot will do the searching for us? Will there be AI bot optimisation techniques we’ll need to learn? If I ask my AI personal assistant, “What’s the most important news story of today?” how will it determine what that is for me, and where will it go looking to deliver me the story?

These are fascinating questions that I think will require creative solutions — ones we are only scratching the surface of at the moment. 2016 is the year when we’re going to have to start figuring all this out! For more on chat apps and bots, read the “Guide to Chat Apps” report published by the Tow Center.

Trushar Barot is mobile editor for the BBC World Service.