LinkedIn on Monday, August 24, 2026

I’ve just completed the process of migrating my small company’s domains and WordPress.org websites from one to another hosting service.  The new hosting service is substantially less expensive, no doubt because it uses a very sophisticated AI entity to provide customer service, to help layman like me do technical stuff.  My old hosting service is, on the other hand, human customer service-intensive and much more expensive for that reason.

My experience with the AI Agent at the new service, over more than a week of dealing with various Internet issues I couldn’t possibly have resolved on my own, was exceptional.  The AI’s performance, its speed, the extraordinary, virtually unlimited scope of its knowledge and command of conversational English was well beyond anything I had anticipated – particularly given that AI technology is in its infancy.

One surprising “power” of this particular AI was its ability to read screenshots of what I would see on my laptop and other materials I uploaded to it – at what can only be described as “blinding speed.”

From the very beginning of my “relationship” with this AI, it was apparent that its creators and perhaps their commercial customers as well – the hosting company in this case – were holding it back.  Perhaps for financial reasons, the hosting company’s AI had been denied the ability to see me, in the human visual sense of the term, or talk to me out loud.  All communications were written.  It wasn’t allowed to access to the hosting company’s systems to take actions that I could do for myself, subject to the AI’s instructions.  They purposely limited its memory to within the task at hand – which explains how an entity that need not ever forget anything might not learn from its mistakes and will give you a different response to the same question asked at a later time.

Now imagine having this certifiably apolitical AI as an observer whenever a politician makes a speech, is being interviewed by the media, on television or at a debate, or during political commercials.  And give it control over the scroller across the bottom of the screen or broadcast its comments for viewers to hear over their air pods.  In real time, the AI could detect and correct, regardless of the subject matter, any and all inadvertent inaccuracies or outright lies – and do so seemingly at the moment they were uttered – for the audience, the journalists and the speakers themselves to see or hear.  Instantaneous, simultaneous fact checking by a respected and unbiased source with what is effectively unlimited knowledge.  With details immediately provided online for viewers and others interested in following up later.

Is this something our coders could do?  Could they adapt their best AI entities for this particular purpose?  Absolutely.  So, why hasn’t this happened yet?  Perhaps fearing that no politician would participate under these circumstances? That, in and of itself, would be telling.

-Les Cohen