Are Twitter bots passing the Turing test?
At Echovo, we’re very interested in the evolution of software and how artificial intelligence is interacting with people (see here). We recently read the Washington Post article, “The curious case of ‘Nicole Mincey,’ the Trump fan who may actually be a bot.” Regardless of where you stand politically, it should get you thinking when the president of United States potentially mistakes a bot for a person. So the question we’re asking ourselves is, “Are Twitter bots passing the Turing test?”
A super brief Turing test primer…
Alan Turing is credited with playing a crucial role in winning World War II by helping crack the enigma machine. He’s also famous for being a founding parent of modern computing.
One of Turing’s interests was artificial intelligence and whether or not machines could think. He’s credited with creating the “Imitation Game,” a way to test this out.
To illustrate, imagine there’s a curtain, on one side sits a human judge and on the other side are a computer and a human. During the test, the judge asks questions. The human and the computer take turns answering. Importantly, the judge doesn’t know who or what is giving the answer. The judge’s job is to determine if the thing giving the answer is the human or the machine. If the judge guesses wrong — that the answer came from a human when it was actually a computer—more often than not then the machine passes the Turing test.
Do Twitter bots pass the Turing test?
So given what we know about the Turing test, do Twitter bots pass? Personally, I’m not sure.
First, Twitter bots are out in the wild and not in a controlled test environment. But there’s probably a clever way to test the theory if you could identify bots and create a case study.
Finally, I’m not sure pass or fail is the right criteria. Thinking, or consciousness, is probably best measured in quantifiable levels rather than a binary system. Michio Kauku explains quantifiable consciousness:
What do you think?
The jury is open for me, but what do you think? Have you ever mistaken a bot for a human only to realize it later? Do you ever go to a Twitter timeline and wonder if it’s a person or a bot?
The final thing I’ll suggest is that it might not matter. Even if you can identify it as a bot, maybe you agree with the bot or appreciate the information and that’s enough.