

I consider these to be the main ethical issues with specifically LLMs and generative AI in general:
- Using people’s work as training data without consent.
- The high cost of training a model meaning that only a few entities in the world can actually do so and so, only few people get to decide what the knowledge base and “slant” of the model is. This is true even for open source models.
- The high resource cost of using a model relative to the value of its output.
- People with malicious intent being empowered by it far, far more than anyone else.
- The model producing the response to the query directly instead of leading to the source, leaving both the source without any way to benefit and the user from having any context queues they can use to verify the reliability of the information.
- Infinite and automated production of misinformation, libel and psychological manipulation.
- Inducing psychosis in people.
Point 1 can be resolved by the people training AI just making different choices. Many won’t unless they’re forced to, but in principle they could.
Points 2 and 3 could hypothetically be resolved in the future with better technology.
The rest are basically inherent to the technology and you can at best try and mostly fail to reduce the risk. So as far I’m concerned, what it would take to build AI ethically is to train it for very specific purposes and have it be used as statistical models by people who know what they’re doing.
Though I do see some potential for ethical LLMs by using them to perform vector searches instead of generating text, basically turning them into smarter search engines.









Subs, always. I don’t really consider dubs an option, unless there’s really, really no alternative. The biggest reason for that is undoubtedly growing up and living in a country where subtitles are the norm for foreign media, but I do think I have some more concrete reasons for it.
Take the line “Hey! I’m walkin’ here!” Probably, you know exactly what I’m talking about and how that line is delivered, even if you’ve never seen Midnight Cowboy (I haven’t). But I can almost assure you that if dubbed in Japanese, it would be thought of as just another line and would never become iconic. If you remove the language and the cultural context it’s spoken in, all that’s left is the literal meaning. And the same is true the other way around, if something is written in Japanese, then it’s going to typically be written with the assumption that it’s going to be spoken in Japanese.
Also, people working on the actual show are going to be involved in the recording in the original language, so it’s typically going to be closer to their intentions. To use an extreme example, Tomino Yoshiyuki (most famous for Gundam) is such a
control freakperfectionist, that you can tell he’s directing from the way the voice actors deliver the lines. But the dubs of all the works he’s worked on over the decades are all going to have different ADR directors with no input from him (and even if he did have input, he probably isn’t fluent enough in English to do so effectively), so there’s no consistent (figurative) “voice”.But most importantly, dubs just make it impossible to do any compensating for losses in translation. There’s a whole spectrum between “This language just sounds like gibberish to me” and “I’m fluent enough in the language to not need a translation”, so as long as you’re not in the former extreme, there are still things you can pick up on: first person pronouns, copulae, honorifics, dialects, levels of formality etc. And especially Japanese is, I think, a pretty translucent-sounding language to people who are used to European languages, so you end up picking up on patterns even if you’re not actively trying to learn it.