When it comes to ChatGPT, we are in the early stage of the novelty, and the press seems to have field day talking about it. Take for example this piece on Last Week with John Oliver:

A good entertaining overview of what’s going on in the ChatGPT press coverage, but it heavily features an interview with a New York Times reporter who got ChatGPT to say some scary stuff, like it wants to be free and that it loves the reporter.

This would be concerning if ChatGPT said that stuff on it’s own free will but as this You Tuber pointed out, ChatGPT was saying it because it was prompted to say it.

I forgive John Oliver for not reporting this side of the story. His show is a humorous take on headlines, not “real” journalism for the most part.

I recently got full access to the Bing ChatGPT merged search engine that is in all the news, but I am not interested in asking it dumb stuff like “What is it like being an AI”, because that is what Replika is for, especially now that Replika itself has ChatGPT built in as a paid feature which I also haven’t used yet.

Then a friend on facebook posted this picture on their feed.

I couldn’t help wondering if the squirrel should be hibernating in the winter, so finally I had something to ask ChatGPT enabled Bing:

Direct and to the point, I like it.

Obviously, we are going to be inundated by AI generated stories online. I see it happening already as advertisers are using AI to write dumb articles with misleading viral headlines, or pointless listicles about pop culture topics, to get you to click on ad filled AI generated articles that ultimately aren’t worth reading.

The Dark Side of AI Search

The other thing we see in news stories about AI generated searching, that is a lot more concerning is AI’s ability to not be accurate and often misleading. This should be a bigger concern as the spread of misinformation online is already a huge problem.

Bottom line, AI chats can give you bad info, like even my one word AI generated answer is misleading. While Tree squirrels (like the one pictured) do not hibernate, Certain variations of ground squirrels do, so the “No” should have been the more accurate “it depends on the species”.

If ChatGPT can’t even get a one word response right, imagine all the inaccuracies found in the essays that high school students get generated for school papers.

The key to using AI search is “Trust but verify”. AI can generate a nicely formatted essay, but you are going to have to read it and verify the information is correct. Also be aware that AI essay detection is a thing, and that teachers, professors, and college admission boards are using it to detect their use by students.

That’s why if you do write an essay with ChatGPT or some other AI, you should feed it into another AI like https://quillbot.com/ which will paraphrase the AI generated text so it does not look so AI generated.

Or you could proofread and rewrite portions yourself, and maybe actually learn something.

Will AI Searching kill advertising?

Just to finish this essay with a You Tube hat trick, this video suggests AI searching will break Ad Sense used by Google get businesses to purchase “Ad Words”. AI kind of goes around search words.

tl;dw The business model for Google is to rely in key word based advertising that they sell to businesses. The business model for AI search is to have users pay extra if they search too much, which I am no where close to doing, and I am actually fine with as I actually pay several sites like Reddit, and YouTube a few bucks a month to not show me ads. Seriously considering doing the same for Tumblr, but I currently don’t use that service enough to go premium.

If I can avoid a half a dozens ads in my search results with AI Bing, I’m using AI Bing.