The philosophical problem of other minds is an epistemic challenge: while I know I have a mind, how do I know if other beings have minds as well? The practical problem of knowing whether another person’s words match what they are thinking also falls under this problem. For example, if someone says they love you, how do you know if they feel that professed love?
Descartes, in his discussion of whether animals have minds, argued that the definitive indicator of having a mind (thinking) is the ability to use true language.
His idea is that if something talks, then it is reasonable to see it as a thinking being. Descartes was careful to distinguish between what mere automated responses and actual talking:
How many different automata or moving machines can be made by the industry of man […] For we can easily understand a machine’s being constituted so that it can utter words, and even emit some responses to action on it of a corporeal kind, which brings about a change in its organs; for instance, if touched in a particular part it may ask what we wish to say to it; if in another part it may exclaim that it is being hurt, and so on. But it never happens that it arranges its speech in various ways, in order to reply appropriately to everything that may be said in its presence, as even the lowest type of man can do.
This Cartesian approach was explicitly applied to machines by Alan Turing in his Turing test. The idea is that if a person cannot distinguish between a human and a computer by engaging in a natural language conversation via text, then the computer would have passed the Turing test.
Not surprisingly, technological advances have resulted in computers that can engage in behavior that appears to involve using language in ways that might pass the test. Over a decade ago IBM’s Watson won at Jeopardy in 2011 and then upped its game by engaging in debate regarding violence and video games. Since Watson, billions have been poured into AI and some claim that AI models can pass the Turing test.
Long ago, in response to Watson, I jokingly suggested a new test to Patrick Lin: the trolling test. In this context, a troll is someone “who sows discord on the Internet by starting arguments or upsetting people, by posting inflammatory, extraneous, or off-topic messages in an online community (such as a forum, chat room, or blog) with the deliberate intent of provoking readers into an emotional response or of otherwise disrupting normal on-topic discussion.”
While trolls are apparently awful people (a hateful blend of Machiavellianism, narcissism, sadism and psychopathy) and trolling is certainly undesirable behavior, the trolling test does seem worth considering.
In the abstract, the test would work like the Turing test but would involve a human troll and a computer attempting to troll. The challenge would be for the computer troll to successfully pass as human troll.
Obviously enough, a computer can easily be programmed to post random provocative comments from a database. However, the real meat (or silicon) of the challenge comes from the computer being able to engage in (ironically) relevant trolling. That is, the computer would need to engage the other commentators in true trolling.
As a controlled test, the trolling computer (“mechatroll”) would “read” and analyze a selected blog post. The post would then be commented on by human participants—some engaging in normal discussion and some engaging in trolling. The mechatroll would then endeavor to troll the human participants (and, for bonus points, to troll the trolls) by analyzing the comments and creating appropriately trollish comments.
Another option is to have an actual live field test. A specific blog site would be selected that is frequented by human trolls and non-trolls. The mechatroll would then endeavor to engage in trolling on that site by analyzing the posts and comments.
In either test scenario, if the mechatroll were able to troll in a way indistinguishable from the human trolls, then it would pass the trolling test.
While “stupid mechatrolling”, such as just posting random hateful and irrelevant comments, is easy, true mechatrolling would be difficult. After all, the mechatroll would need to be able to analyze the original posts and comments to determine the subjects and the direction of the discussion. The mechatroll would then need to make comments that would be trollishly relevant and this would require selecting those that would be indistinguishable from those generated by a narcissistic, Machiavellian, psychopathic, and sadistic human.
Years ago, I thought that creating a mechatroll might be an interesting project because modeling such behavior could provide useful insights into human trolls and the traits that make them trolls. As far as a practical application, such a system could have been developed into a troll-filter to help control the troll population. I’m confident that the current LLMs could engage in trolling with the proper prompts, although they would lack the true soul of the troll.
A Philosopher’s Blog is Now on Substack!
You can subscribe and read for free.