Chatbot Ethics: Does It Pay to Say Thank You to a Bot?
The question of whether or not we say thank you to a chatbot -- or if we know we're talking to one -- reveals some of the ethics around bot technology.
By Edward Pollitt
April 23, 2020
You thank your Uber driver. You thank your waitstaff. You thank a shop assistant. But do you say “thank you” to a chatbot?
The bot doesn’t care if you don’t. After all, the bot isn’t human. But as more and more services become automated, will saying “thank you” no longer be the norm? It’s not necessarily just an issue of manners; rather, it shows some of the wider ethical issues that chatbots bring with them.
The first thing to understand, when looking at the interplay between our innately human sense of manners and the AI-powered chatbot, is that not all bots are created equally.
In recent years, we’ve seen bots and bot technology deployed pervasively in customer-facing functions—a shift which has, in part, been normalized by the development of devices like the Amazon Echo and other Amazon Alexa devices.
When we visit a new website for the first time, often as a potential customer, we’re regularly greeted by a bot. But here the variations between different bots become apparent.
Some bots have a very limited functionality. In such cases, the bot will simply be able to repeat some of the site’s FAQs and direct a customer’s inquiry to the appropriate (human) customer service representative.
This still provides a great service to the customer and allows the business to use the bot as a first filter for customer inquiries. Importantly, in these instances, it’s very clear to the potential customer that they’re speaking with a bot, not a human. In such cases, it’s fair to assume that a “please” or a “thank you” would be less likely than in a traditional human-to-human interaction.
But some more advanced bots are developed using complex algorithms in an effort to imitate—as closely as possible—the way a human speaks. These ultra-realistic bots are not hidden in the corners of the dark web—they’re out in the open. To see how lifelike a bot can be, take some time to check out Mitsuku.
The business case
When businesses deploy these more complex bots in a commercial setting, the ethical questions become slightly more pointed.
First, we need to ask whether customers have the right to know that they’re talking to a bot.
Often, bots are given a typically human name to emphasize their ability to replicate human behaviors. And with millions of dollars being invested in this technology, it’s safe to assume that there will be some cases in which people aren’t sure whether they’re speaking with a bot or a human. In these cases, it would be more likely for the bot to receive a “thank you.”
So what are the ethical ramifications if someone is speaking to a bot but thinks it’s a human? What if they fall in love?
This conundrum is part of the reason why, in 2018, the State of California implemented regulations that require companies to disclose whether they’re using a bot to communicate with the public. Disclosure can be something as simple as a quick, “Hi, I’m a bot!” displayed somewhere on the bot’s interface.
As well as clearing up ethical gray areas, a bot disclosing its automated identity also assists the level of service provided in the interaction.
Customers are more likely to be direct and to the point when they know they’re dealing with a bot, as opposed to worrying about the emotional impacts of their question. In other words, the customer will be less likely to stop and think, “Am I being rude?” as we often do in business and retail environments.
And isn’t this what automation is all about—removing the human factors that slow us down and replacing them with data-driven solutions?
But the way we deal with bots today won’t necessarily be effective for long. An article in SmallBizGenius predicts that bots will account for 85 percent of all customer interactions by 2021.
There’s also the possibility that the current COVID-19 outbreak—and the associated self-isolation and quarantining—will accelerate the adoption of bots.
So how can our manners and ethics keep up with the normalization of chatbots? Google Assistant is already trying to improve the conversation between bots and users by being “nicer” to users who say “please” and “thank you”—kind of like people do in real life.
As chatbots continue to improve, it’s likely that human users will continue to say “thank you” more and more. This isn’t just a result of initiatives such as that of Google Assistant; rather, it’s because chatbot technology will continue to improve to the point where the quality of these conversations so closely replicates human-to-human interaction that it will be second nature to exchange such pleasantries.
So remember, always thank your chatbot.