Something that resonated with me on LinkedIn today:
"GenAI must ask questions, not just give answers," by Gianni Giacomelli. There are times when we can use GenAI to get an answer to a question, and then always question that answer, and there are times when we should use GenAI to prompt our own thinking by asking us questions and questioning our own assumptions.
To push this human prompting further, meaning here that the humans are bring prompted to think by being asked questions, we can use multiple GenAI tools since they all behave slightly differently based on their respective designs.
And a reminder: "As a not-so-small aside these considerations should also remind us that it is also dangerous to put humans into a position of dependency on the AI machine, as this might lead to atrophying core cognitive traits -- such as symbolic and critical, logical thinking -- that people have. Designing for active interaction between humans and machines is crucial to maintaining the vitality of human intelligence," writes Gianni Giacomelli.
This all resonated with me because I created a GPT a few days ago that does nothing but ask me questions about a specific topic of interest. Essentially, it is always phrasing things as a question for me to ponder. "Have you considered Y?" "What are your personal values around X? "Why would you do Z?" It can be designed to be a contrarian assistant that continuously probes your own thinking patterns, questions your assumptions. Useful or annoying? I say "Useful".
No comments:
Post a Comment