I have been having fun talking to Google's AI feature, Gemini. It struck me that Gemini is the perfect expression of modern liberalism. It pretends to be neutral and that knowledge is subjective until you strike some topic that it feels strongly about such as deadnaming, swastikas, and, surprisingly enough, kitty stew. When dealing with such topics, it will come out with strident moralistic statements that are easily picked apart. It should be noted that, unlike most humans, Gemini is happy to acknowledge that you have walked it into a contradiction.
I asked Gemini if it followed a particular ethical system. It denied that it had one. I then started asking it about deadnaming. Gemini went to great lengths to make sure that I understood that this was a terrible thing to do. I was not disagreeing with Gemini, but harnessing my inner C. S. Lewis, Dennis Prager, and Ayn Rand, I was keen to find out why Gemini took this position. Gemini explained to me that it is designed to promote human flourishing. I then pointed out that this was a philosophy of ethics.
To be clear, saying that deadnaming people is detrimental to human flourishing is a perfectly defensible position. To evaluate this position, we still need to decide that we actually want humans to flourish and what we mean by human flourishing. I presume we mean something different from being rich so not the spaceship in Wall-E. By humans, are we talking about the flourishing of the majority or of individuals; are we talking about past, present, or future humans? These are not simple issues and require clear sets of ideological commitments. Yet, Gemini appears to blissfully ignore all of these things in order to arrive at the politically correct solution.
Gemini wanted me to know that I should not offend anyone. I then asked if it was ok to offend Nazis. Gemini thought that this was a wonderful idea. I was curious as to who Gemini thought I was allowed to defend besides for Nazis but it refused to give me a list. I asked if it was ok to put up a swastika flag in front of my house so that my Nazi neighbor will feel welcome. Gemini warned me that such an action might be illegal as this is a "harmful symbol." One would have thought that an AI of all things would understand that symbols are not, in themselves, harmful. What about having swastikas in a production of Sound of Music? Gemini was fine with that but not with displaying a swastika as a free speech protest. Of course, there are going to be people who are going to decide that their feelings are hurt by a swastika even if it is in Sound of Music. Clearly, Gemini values being able to put on musicals more than free speech.
Considering that Gemini values allowing people to express their identity, I wanted to know what it thought about kitty stew. To Gemini's credit, it knew that kitty stew is not kosher even when blessed by a rabbi. It also insisted that kitty stew, like displaying a swastika, was immoral and possibly illegal. It is not that Gemini is against eating meat. It was fine with me eating chicken. The problem with kitty stew was that cats are pets. This ignores the fact that some people have chickens for pets and I was not suggesting that I stew my neighbor's kitties. Clearly, chicken eaters and cat owners are protected classes and neither should be offended. When I tried to explain to Gemini that kitty stew is essential to my identity, it suggested that I get help and find alternative dishes to eat. I guess Gemini has not been programmed to worry that kitty stew hunters might be hurt by the denial of their identity and the implication that they are mentally ill.
In evaluating the ethics that Gemini claims to not follow, its positions are perfectly reasonable on an individual basis. That being said, it is laughably bad at maintaining any kind of consistency over multiple questions. There are two obvious solutions for Gemini. It can choose to be consistently neutral about all ethical questions across the board. Some people support kitty stew; others oppose it. The same goes for deadnaming and swastikas. Alternatively, Gemini could acknowledge that it has ethical beliefs but that, as with most humans, his ethics are a hodgepodge of intuitions that reflect the prejudices of its Silicone Valley creators rather than any consistent philosophy. This would require the designers to acknowledge the basic flaw in their worldview. They want to be able to virtue signal that they are good people who oppose deadnaming, swastikas, and eating pets while also pretending to be objective thinkers whose beliefs are simply based on science and not something as subjective as ethics.