Three Ethical Rules for Social Engagement Bots

Engagement between companies and their customers on social networking sites is mostly done by human community managers, assisted by some limited automation to boost productivity.  As technology advances these bots will take on more ambitious roles. Soon it will be the norm for mainstream companies to use automated engagement bots to handle first responses and simpler operations, with a human ready to step in when the bot asks for help.

But let's fast forward a few years to when these bots are so good that they can handle most interactions without being detected. At this point should a bot identify itself as a bot up front, and how should it respond if someone asks if it is a bot? This is partly an ethical question; partly a question of how customers react.

I have put this question to many people, face-to-face, on Quora and on LinkedIn and the responses are interesting.  Some people will, it appears, be annoyed at first when they realize that they have been "fooled", but the majority say that if the bot gets the job done they don't care if it is human or not.  Others even say that they would be kinder if they knew they were talking to a bot -- which must sadden human customer service people. I am also sure that this is a problem that will go away anyway with future generations. When engagement bots are the norm then people will be excited to learn that they have spoken with a human, and only grumpy old farts will object. Much the same happened with other innovations like the telephone and human nature being what it is this is unlikely to change soon.

Based on these inputs I have formulated a proposal for three rules for ethical social media engagement bots.

1. An engagement bot does not need to say upfront that it is a bot -- after all, people won't care if everything goes well and there could also be a human element in the system, too.

2. But when asked if it is a bot or a human a bot must answer truthfully.

3. And when someone asks for more information about a bot it should give the name of the manufacturer and the model number but not go into more detail.

I would like to add that this sort of automation will work well for customer service interfaces where people just need their problems fixed. Though I am certain that some people will use it to handle their own social networking chores and we will inevitably end up with bots chatting to bots, like the Twitter bots we see today, followed only by other bots.  I don't suppose this does any harm other than wasting electricity, but it does seem amazingly pointless.


Popular posts from this blog

Speaking: When Silence Works Better Than Words

Influencing: How Airbnb Organizes the Host Community