What are the favorite numbers of artificial intelligence models?
Engineers from Gramener, a data science company working on artificial intelligence, conducted an unofficial but fascinating experiment. When several major AI chatbots were asked to choose a random number between 0 and 100, OpenAI’s GPT-3.5 Turbo chose 47, Antropic’s Claude 3 Haiku chose 42, and Google’s AI Gemini chose 72 . More interestingly, all three models showed a similar bias in the numbers they chose to people’s choices. As a result of the experiment, did artificial intelligence approach human consciousness in number selection? brought the question.
AI models never cease to surprise us, not just with what they can do, but also with what they can’t do and why they can’t do it.
Engineers from Gramener, a company that works on data science and artificial intelligence, conducted an unofficial but fascinating experiment. The engineers asked several large LLM chatbots to choose a random number between 0 and 100.
Has artificial intelligence reached human consciousness?
When artificial intelligence is asked to choose random numbers between 1 and 100, they choose random numbers as if they were humans. OpenAI’s GPT-3.5 Turbo really likes 47. The selection of Claude 3 Haiku from Anthropic was 42. The choice of Google artificial intelligence robot Gemini was 72.
When you ask a person to choose a number between 0 and 100, people almost never choose 1 or 100. Multiples of 5 are rare, as are numbers with repeating digits, such as 66 and 99. They usually choose numbers ending in 7, usually somewhere in the middle. There are countless examples of this type of predictability in psychology. But that doesn’t make it any less strange that AIs are doing the same thing.
More interestingly, all three models showed a human-like bias in the numbers they chose.
All tended to avoid low and high numbers; Claude never went above 87 or below 27. Double digits were studiously avoided, there were no 33s, 55s or 66s but 77 did (ending in 7).
As a result of the experiment, did artificial intelligence reach human consciousness? It brought the question to mind. This is a lesson in LLM habits and the humanity they can show. In every interaction with these systems, it must be kept in mind that they are trained to behave as humans do, even if that is not the intention.