In 2016, Facebook and Microsoft opened up new platforms so software developers could connect chat gadgets to their consumer websites and apps. These gadgets would start a conversation and effectively direct the consumer to the right service, or even solve their problems. Google quickly followed suit. We had already been pleasantly surprised by what Google’s Assistant and Amazon’s Alexa could do. And we already knew that people could chat for hours with a computer.
Chatbot Hell is a more subtle story than you might think, and is not just a parable of how technology is oversold, but how managers use technology to hide their own failures.
In 1966, a computer science professor at MIT, Joseph Wiezenbaum, created a foul language generator that mimicked a psychologist. When you typed something in Eliza, she repeated what you typed, usually in the form of a question. But Wiezenbaum was appalled to find that people were shedding their intimacy with Eliza and spending hours in such sessions. Media historian Dr Simone Natale, author of the book Misleading media: artificial intelligence and social life after the Turing testoffered an excellent description of communicative AI like Eliza, Siri, and chatbots: a “banal deception”.
A few years later, Google came up with an amazing AI demo in which a bot made an appointment with the hairdresser, without the hairdresser realizing he was talking to a machine.
“A machine playing the role of a human? Well, someone move the doomsday clock forward, the singularity is almost here,” wrote a very excited ScienceAlert reporter.
Today, that promise remains unfulfilled.
Hopes far exceeded what could truly be achieved, and when bots went wrong, mistakes could be catastrophic. Former Health Secretary Matt Hancock often raved in media interviews about a chatbot used by the GP at Hand app, but clinicians feared he was giving misleading advice , like not recognizing a heart attack.
Worse still, a French remote healthcare company that was evaluating a chatbot posing as a suicidal patient was told, “I can help with that.”
Don’t expect miracles, a CTO experienced in implementing chatbots tells me: “This is where web design was in 1996, and most companies don’t know they need a web designer.”
He argues that a well-designed chatbot can still quickly triage customers, a boon for small businesses that can’t afford a call center. But only if the company works with the limits of technology.
Broadhurst agrees, citing Amazon’s highly efficient returns process.
Too often, however, companies that run sprawling empires of chaotic customer processes are content to apply new technologies to them, which is like putting a fresh coat of paint on a dilapidated house.
“It’s a vicious cycle. Companies under pressure use AI to push people away,” says Broadhurst.
Outsourced call center staff are already on a leash, often unable to escalate issues, and deprived of useful knowledge about the business. A bot is not going to improve this. “If the human can’t help, then don’t involve the bot.”
Technology cannot adequately replace human contact when issues require sensitivity, discretion and judgment, says Jo Causon, CEO of the Institute of Customer Service. I would tend to agree – which doesn’t make someone a Luddite, just a techno-realist.