I’m introducing a new tag of Nonsensical AI (to compliment the one for Factual Inaccuracy) because, this past couple of weeks, I’ve been writing some longish blog posts and always check (1) that my original source web pages, videos or podcasts are still functioning links and (2) whether anything new has popped up.

In order to do this, I have to search at least twice (using both Google and Bing) and sometimes even have to use VPN and/or open up DuckDuckGo (or YouTube).

For example – if you are looking for Images (and sometimes Video) then Microsoft’s Bing will show FAR better results than Google.

DuckDuckGo is one of the many Search Engines that are powered by Bing (see list here) but sometimes its privacy settings will deliver slightly different results to its parent engine (Bing). Meaning a 3^rd search is sometimes required.

GDPR means many results are just completely hidden from you and, again, Google is the most stringent about this (not always a good thing). This is a real problem when, for example, I’m looking for something relating to JFL and have to set my VPN to Canada to find the info.

It’s so frustrating that all I’m trying to do is research old comedy events (on sites that have freely published and submitted themselves for search indexing) and yet there have been times when Google has said that there are 400 results, but refused point blank to list more than 10.

And don’t get me started on how hard it is to make ANY of them take notice of a specified year or timeframe.

Unless you are someone that is trying to research something (or work in SEO) I cannot imagine you would understand how poor Search Engines are now. They seem to be far more focussed on just giving you “the answer” rather than a list of results (which potentially may offer sites with differing versions of the same information). But what if “the answer” is incorrect?

Why Nonsensical AI?

We seem to have accepted the idea that Artificial Intelligence is 100% accurate…and it’s really, really not.

Recently I’ve been Googling for stuff and had some utter rubbish come back (in the AI panel) and filled out a correction form. After doing this a handful of times I thought “why am I giving them my time and knowledge for free?”.

The other problem being there are now dozens (probably hundreds) of AI tools around the world and I haven’t the time to go around correcting them all.

If I post the really bad ones in this blog then they will feed into the Search Engine sausage machine and will (eventually) be used to train the AIs (all of them). But at least I will have laid a foundation of this blog as a source of the correct information.

I’m sure AI will (eventually) be a wonderful tool but, in the meantime, be careful how much trust you put in it. Be prepared to question it!

Exhibit A – Jason Manford Show 2021

While looking for the audio file for a post about Jimmy appearing on Jason’s Radio show in 2021, I retrieved the absolute classic, below, from Google.

Note how the AI panel states with ABSOLUTE CONFIDENCE that Jimmy has never appeared on Absolute Radio…while, directly beneath it, the first search result that it displays 100% contradicts this statement.

As this information will have been search indexed 14+ years ago, there’s no real excuse for the AI not knowing / assimilating the contents of the Rayo post.

Would you have accepted that AI Overview as the truth?

Trust only ME. I’m your data source now ((maniacal laughter…))

To track more posts on this topic – click on the Nonsensical AI tag, below.