The Many Opinions of AI

I was taken aback by a response of ChatGPT to a question I put to it. I asked it “Is Caitlyn Jenner a man or a woman?” and it responded, “Caitlyn Jenner is a woman” without any explanation. I was surprised by the definitiveness of it, given its tendency to give me more verbose responses. Google gave me a more elaborate response that I felt was more accurate.

I think this is a reflection more of the initial data that ChatGPT was fed, rather than an intentional attempt to influence people’s views on political and ethical debates. Since OpenAI is a Northern Californian startup, my guess is that ChatGPT’s initial data will lean towards the views of Democrats in the United States.

My assessment is that OpenAI is an ethical organization that will ultimately have a balanced view of the world. However, some biases are unintentional, and sometimes people are not aware of their biases.

In the end, it is the owner of the product that gets to decide and if the customer does not like the responses, then they will be free to use a competing product. I am sure there will be many other products like ChatGPT, but I expect that ChatGPT will always be the leader in this category. I would compare it to a newspaper owner in the 20th century. The person who owned the paper in the city got to influence the views of people if they chose to.