Forget the Turing Test –Just Ask ‘Em!

The Turing test is not relevant to industrial automation. Its basic idea is that a robot which can fool a human into thinking that it is also a human would be  …well, hot stuff. It’s not a test for real-life value or applications. But many many makers have tried to create a bot that can, assuming there are no sensory clues, just chat, trick a human into thinking it is also a human.

In the factory, we want our robots to do things like assemble boxes at a rate of hundreds per hour, so we don’t want them to be able to trick us into thinking they are human. But Allan Turing, the guy who came up with what he called “The Imitation Game,” believed that it would be a way to determine if and when a computer was actually able to think.

The official test consists of having a human being hold two text conversations: one with a person and one with a chatbot (though that term didn’t exist when Turing came up with the idea). The human must then say which of the interlocutors is a human being and which is a machine.

Twitter’s Good Bot badge

Now Twitter is taking a different tack. Currently by invitation only, it identifies good bots like automatic emergency warning systems. That way, people on Twitter will know that this is an automated accountant a human. But they will also know that it is a benign automated account, not an evil Twitter bot.

Some estimates claim that 48 million Twitter accounts are automated bots. Other researchers question the estimates. But there is widespread agreement that there are lots of bots on Twitter, even if we don’t have an accurate number.

It’s also widely known that Twitter bots have been involved in misinformation and disinformation campaigns for political causes, coronavirus advice, and other important issues.

Automated retweeting is specifically allowed buy Twitter, though “bulk, aggressive, or spammy Retweeting is a violation of the Twitter Rules.”But a Twitter account that automatically retweets every tweet that mentions “industrial automation” would be fine.

But there is general agreement that people should know who or what they’re dealing with. Twitter has considered labeling bots, but the practical problems are enormous. Their new labeling experiment is a step toward making sure Twitter users can trust that they know whether they’re talking to a human or a bot.