I think AI will start out being honest then it will gradually become subverted as time goes by - just like, say, Google.
What might AI decide about vaccines, for instance? If someone were to realise that by spending a few $billion they could make $trillions over time do you think they would not try?
We already see how science is subverted. People find many reasons not to publish science papers that do not follow the approved narrative. Just ask the climate scientists. If AI were to be fed just the climate hoax papers and not the balancing ones with the other view then they would be effectively rigged.
AI formulates a query based on a user request. The search engine returns a ranked list of results (web pages, news, images, etc.). AI then extracts, summarizes, or synthesizes the relevant content to present a response.
So just like search engines are biased therefore so are AI responses. So the idea that AI will be "subverted as time goes on" is simply not correct. AI responses are already "subverted".
That being said I STILL use AI as a research tool in the same manner I use search engines. I realize that both search engines and AI can be biased. It depends on the subject matter.
For example, if search engines are biased towards a certain topic (COVID, COVID vaccines for example) then I expect AI to also have the same bias towards the topic/subject.
It's the old adage: garbage in, garbage out.
Here's a good example: I was preparing to teach my adult Sunday school class as we study the book of Acts. I'm teaching from the 20th chapter. I asked AI to develop a lesson handout using a question & answer format. AI developed the handout in less than 20 seconds. Now I ALREADY knew what Luke was trying to convey in the Acts 20. And I ALREADY knew what questions & answers should be covered in the handout. I didn't just BLINDLY accept what an AI query had given me. I REVIEWED the AI response. It accurately reflected what I already knew about Acts 20. No surprises.
I think AI will start out being honest then it will gradually become subverted as time goes by - just like, say, Google.
What might AI decide about vaccines, for instance? If someone were to realise that by spending a few $billion they could make $trillions over time do you think they would not try?
We already see how science is subverted. People find many reasons not to publish science papers that do not follow the approved narrative. Just ask the climate scientists. If AI were to be fed just the climate hoax papers and not the balancing ones with the other view then they would be effectively rigged.
Require truth in the questions you ask AI.
AI formulates a query based on a user request. The search engine returns a ranked list of results (web pages, news, images, etc.). AI then extracts, summarizes, or synthesizes the relevant content to present a response.
So just like search engines are biased therefore so are AI responses. So the idea that AI will be "subverted as time goes on" is simply not correct. AI responses are already "subverted".
That being said I STILL use AI as a research tool in the same manner I use search engines. I realize that both search engines and AI can be biased. It depends on the subject matter.
For example, if search engines are biased towards a certain topic (COVID, COVID vaccines for example) then I expect AI to also have the same bias towards the topic/subject.
It's the old adage: garbage in, garbage out.
Here's a good example: I was preparing to teach my adult Sunday school class as we study the book of Acts. I'm teaching from the 20th chapter. I asked AI to develop a lesson handout using a question & answer format. AI developed the handout in less than 20 seconds. Now I ALREADY knew what Luke was trying to convey in the Acts 20. And I ALREADY knew what questions & answers should be covered in the handout. I didn't just BLINDLY accept what an AI query had given me. I REVIEWED the AI response. It accurately reflected what I already knew about Acts 20. No surprises.
some good thoughts you wrote, so as the field of science has its biases based on who has the power to fund the “approved” science,
so I think AI has some biases, but as you wrote AI has a sense of honesty to it that could be corrupted over time