Chatbots and voice bots are terms for computer programs that people interact with by talking to them, either through written messages or spoken word. Probably the most famous of all such bots is Amazon’s Alexa, which will respond with useful replies if spoken to with commands like, “Is it going to rain?” or “What is this song?” Most of the big technology companies have equivalents to Alexa, including Apple’s Siri and Google’s Assistant.
To date, text-based chatbots of the kind that a user might talk to over Facebook Messenger or WeChat are facing a backlash. Having been promoted as a possible significant paradigm shift for computer interfaces, they do not appear to have disrupted major businesses as when shopping shifted from retail stores to the web. From an investor’s perspective, disappointingly few breakthrough businesses have used the conversational format to break open a new market. In the words of Digit’s Ethan Bloch (Asay 2018), “I’m not even sure if we can say ‘chatbots are dead’ because I don’t even know if they were ever alive.”
At the same time as bots are being pronounced dead, a different kind of conversational AI bot is being criticized for being entirely too alive. These are bots used to post content on social media platforms, pretending to be humans with particular political or ideological beliefs. The purposes of these bots vary from straightforward campaigning and message amplification, through to much more sophisticated and cynical confusion generation — where bots are employed to pollute the public debate with so much confusion that no form of consensus on desirable actions can form. Social media companies are aware they are being used to disrupt public debates and the conduct of governments, and are fighting an ongoing arms race with bot creators to reduce this disruptive noise.
It is not yet obvious who will win, the bot writers or the platforms that host and fight them. In Twitter alone, out of its 336 million users worldwide so far, researchers estimate up to 50 million are bots (Varol et al. 2017). Furthermore, these bots are prolific communicators. A recent Pew Research Center study suggests that up to 66 percent of links shared on Twitter come from suspected bots (Wojcik et al. 2018). And, the problems are not limited to bots that live on social media platforms. For instance, researchers found that over 5.8 million submissions made to the U.S. Federal Communications Commission on the topic of internet neutrality were fakes produced by bots (Flaherty 2017).
Bots with social and political objectives now get such a bad press that it is worth noting they are also sometimes used for more unambiguously positive purposes, such as increasing job opportunities in Ghana, combating sexual abuse in Liberia, and facilitating access to welfare programs in India and the Philippines. Outside of social media platforms, they are used by thousands of online businesses to offer helpful chat-to-us services on company websites. These pop-ups often give customers access to what they want faster than simply clicking around. 17These days a phone call to a bank will be partly or entirely handled by a voice bot to answer questions like “What is my balance” more quickly and cheaply than a human can. While AI chatbots might not fundamentally reshape the economy, they do matter, and unlike many new digital technologies chatbots may actually matter more to civil society than they do to business.
There are two reasons why bots will continue to have a role in driving greater citizen engagement in both activism and government affairs.
The first is that bots provide one of the only scalable ways of communicating through popular instant messaging tools with large numbers of supporters. A campaign that has acquired permission to send personal message to thousands of users of a chat tool like WhatsApp or Signal will get more of those users attention than almost any other current way of communicating. This means that as much as chat platforms try to limit unwanted or distracting content, activists will keep trying to find ways round this.
The second reason that chatbots have a future is that the coordination of individuals who do express an interest is complicated and potentially chaotic. With a bot asking simple questions like “What day are you free?” or “Can you help with this task?,” a certain amount of administrative tasks can be reduced, freeing up campaigners to do work that really demands human intelligence.
Bots may be increasingly used to support coordination and collective action in a seamless and incremental manner. Based on the data they collect, and drawing insights from machine learning and predictive analytics, bots will become better and better at connecting citizens who are most likely to engage in activities together and who have complementary skills for collective action, and at targeting requests for action (e.g., donate, attend a rally) based on the data they collect.
17: The business use of “Any questions?” chatbots on company websites are usually human-bot hybrids. The bot is used to give a quick and friendly response while a real operator becomes available. ↩