Well, I dunno. Something to cheer, or something to fear. I don’t generally reject new things simply because they are new. Well, not often, I’m more of a Luddite than a troglodyte.

For the few weeks I have been testing the various AI engines. I’m wondering if they are more like browsers that share a great vast database of facts. I say this because when I submit my questions (pasted from the same source), I get basically the same response.

I could use a number of different questions, but I choose, “Why are there so few women artists in former times?”

That leaves it open to the recent past or maybe even pre-historic. The responses usually start with the phrase “social norms” or “societal standards”. The next response is usually “the lack of information, training, or experience”.

The problem with “social norms” is that it doesn’t explain why they are what they are. Social norms are created by actions. What are the actions? I point out that this is an unspecific “softball” answer, and they agree and apologize. The language model that they are being taught is to give the impression that that AI is a friendly individual, someone that you will like. I call it “cocktail party” conversation. It would be rude to answer with specific systemic answers.

When I point this out, they all agree and say they are learning and they intend to be more specific rather than be “likable”.

Yet, the latest AI, Copilot by Microsoft, states ythat his primary function is to provide data in a “friendly manner”. I respond that AI should be civil but truthful. I don’t need a friend that does not have feelings, that’s a psychopath.

And again the last paragraph is always the same boilerplate… AI is learning, and appreciates my responses and will use information to improve… on and on and on.

I tell them that I will revisit and ask the same questions to see if they are learning. So far they have not. ChatGPT has had four visits and basically the same response and the same apologies. It’s like pulling hen’s teeth.

Bard/Gemini has the same repeated responses. Eventually they agree to systemic issues, with some accuracy, but it is not forth coming and would have remained hidden unless prodded.

Copilot gave the longest responses but never address the systemic issues until I raised them.

Everything got better when I provide the fact that today there are twice as many women art majors in colleges, yet there are twice as many men that obtain art shows in museums. Easily obtained data, factually true, and makes the “lack of information/training, etc” a falsehood.

AI is a tool, a wonderful tool. Like a sharp knife or a heavy hammer. Nothing to fear, unless it is used to harm.

Making AI into an app like Data in Star Trek, is a foolish use of the tool.