Member-only story
ORIGINAL RESEARCH IN DATA SCIENCE (PILOT STUDY)
Why AI loves “M” when asked to pick a random letter
And what this pseudo-randomness tells us about AI determinism
As my research is starting to get formalized academically, if you are a fellow researcher replicating or investigating this further, please reference my work:
Nightingale, J. (2025, February 26). “Why AI loves “M” when asked to pick a random letter.” Jim the AI Whisperer. https://medium.com/@JimTheAIWhisperer/why-ai-keeps-choosing-the-letter-m-6ce3a0990f49
AI models show bias when selecting a “random” letter, often favoring M. This reveals AI’s deterministic nature.
Today I have results of another experiment in AI to share with you (yes, I know; I’m making discoveries almost faster than I can document them)!
If we ask a Large Language Model (LLM) to randomly select a letter of the alphabet, we can usually correctly guess which letter the AI will choose. It’s not a magic trick! It’s a quirk of the inherently deterministic patterns of AI.
I’ve found there’s an inbuilt bias for models to favor certain letters while giving the illusion of random selection…
