4

  • Would love to see a blog post about how Kagi views the environmental impact of AI and the ethics of how LLMs are trained

  • SuggestionsDesktopMobileAIDone

Sorry if this isn't the right place to post this, since it's not strictly a feature, more of an info request / blog post request.

There are three big concerns amongst my friends and I, when it comes to the use of LLM generative AI:

  1. The environmental impact. Training LLMs takes huge amounts of water and energy. Is Kagi concerned about the environmental impact these models have, and are they taking any steps to do anything about it?

  2. The ethics of the training data LLMs use. A lot of LLMs are trained on content created by humans, essentially with the intent to profit off of it without compensating those humans accordingly. Additionally there are some concerns about the usage of data as training material without explicit consent. What are Kagi's views on this, are they concerned about the ethics of training LLMs on copyrighted and other works?

  3. Another side of the ethics of training data used by LLMs is that there is a greater and greater source of LLM-generated content on the web. This both tends to take exposure from original sources (without crediting then and denying them any revenue they might see from it) and leads to a severe degradation of the performance of the LLMs. Is Kagi concerned about preserving due credit where it's due, and are they continuously monitoring the quality of the results their LLMs return?

I believe these are big concerns when looking at Kagi as specifically a more ethical alternative to other search engines, and being able to have these directly addressed would make it a lot easier to feel like Kagi isn't just looking past the real problems with AI because it has its origins in it.

Thanks for taking the time to read this, I hope you take it into consideration.

  • Vlad replied to this.

    This sounds like ideological blackmail masquerading as ethical concern, the kind of concern where a company then is offered a certificate to prove how "ethical" they are in exchange for a large consultancy fee to the "concerned" parties. This might not be your aim, but the way you are arguing and the kind of company you are hunting is a carbon copy of how these groups operate.

      fracturefalcon Hey there

      Will try to answer this in as pragramtic and practical way possible.

      1/2. Kagi is not training LLMs. But if we did, environmental impact of that would probably be neglible to the environmental impact of cars we drive, phones we use, trash we throw or air travel we take. This is to say there is nothing you can do today that does not have enviromental impact and LLMs are the least of our issues as species. I do believe that Kagi as a company and its mission have a net positive effect on the world and that is what matters.

      3/ We believe we will be very suceessful at combatting LLM generated content, because most of it exists for the purpose of monetizing with ads, which is something we are very successful of detecting and downranking in our results.

          No one is typing