So, I am sure most here know there has been a massive uptick in articles and videos where a person asks AI about some current topic. 

AI, whichever one it is they ask, some ask them all to see if they give different answers, then spits out an answer with explanations of what it "believes" is going on. Now, I put "believes" in quotes because I am not sure that you can say AI believes anything. It is a compiler. It compiles known data from the internet, and then forms an answer based on the known data. 

Now here is an obvious giant problem with that, and we have actually seen AI demonstrate this problem, much of the data on the internet is false and/or intentionally misleading. So, if it collects all the "data" it can find to formulate an "opinion" much of the data it uses is data that should have been tossed out. I don't think AI has a good filter for this. I mean, yes, you can ask it to filter out certain types of data. You could ask it to say filter out any idea or opinions expressed by White Nationalists. Still, what you are going to get, is some sort of compilation of information and probably to get any sort of accurate answer, meaning an answer that hones close to truth or fact, you would probably need to include many, many filters. 

I've now watched and read several "I ask AI" articles and videos. I mainly find them problematic. Particularly if the person or "prompter" is asking AI for an "opinion" without inserting any filters. How does AI cope with a person like Trump, that states falsehoods on a regular basis as if they are facts. How high or low does it rank a statement from Trump, the president of the United States?

Now, where AI could be helpful is when you feed into it a fixed set of data, data that you know is true and is fact, and then AI evaluates that data. As an example, I read an article where someone had AI evaluate the Big Beautiful Bill, so the data set fed into AI to evaluate was just the bill. I think it took the AI about 20 minutes to read and evaluate the bill and then encapsulate it into understandable shorter bullet points. OK, that is helpful, particularly to people like us, trying to discover what the bill is, what it does, and how it will impact us without having to read 9000 pages of it ourselves. 

Now, that is essentially a simple task, not an opinion, it would just be a breakdown of a fixed set of data. So, give AI an accurate fixed set of data to breakdown, great you have something there. 

Give AI access to the internet and all that is on it...well...what you have is a mess. 

I am writing this because David mentioned AI and the idea that we must unite. Yes, as people we do need to come together. Not fight over if we are Democrats or Republicans, conservatives or liberals. We need to agree on what is real and true. Here, we are failing as a society and the result of that failure will be bad things are going to happen to us. 

It does strike me that Trump's main goal, the biggest thing he attempts to do, is to contradict reality. To put out as much false information as he can. Is this an intentional effort with the specific purpose of impacting the kind of answers AI gives or comes up with?