AI slop & blockchaining attention

I've heard concerns about misinformation becoming a significantly harder problem as AI slop proliferates.

 

Initially I reasoned that humans generally care about quality information and would simply choose better sources in order to sidestep the garbage; but a cursory look at present news consumption habits (e.g. sourcing from social media and messaging apps, rather than credible sources like The Economist or more trustworthy institutions like academia) doesn't presage well for this opt-in mitigation.

However, there will for various reasons likely always be a significant demand for credible information. Which makes me wonder about the foundations of credible information: typically experts' communications and research, I reckon. Which are typically performed and published by humans, even when assisted by technology. 

Being able to verify that a claim was made, or data sourced, from an expert, might then become critical and valued. I can imagine a device/technology that constantly accounts for an expert's attention in, say, 15 minute increments, and that accounting is blockchained for reliability. Future claims must be attached to some not-already-claimed portion of that expert's attention- since it is only the spending of an expert's attention that produces the constituents of good information. Deep fakes of celebrities might also be rebutted in this way: exactly when did that celebrity say whatever the video claims? Sorry, the celeb was verified to have been spending their attention otherwise (e.g. sleeping) at that time. 

A precedent for this is lawyers' timekeeping: they often account for billable hours to the nearest 6-minute increment. 

I can further imagine a third-party verification provider, who reviews claims against a given individual's attention blockchain, and replies with either 1) consistent with or 2) not consistent with, the claim. This would allow for privacy (the inquirer doesn't know what the celeb was doing during that time, just that they weren't recording a video). Think of it as a 24/7 alibi service, whose data the individual owns and chooses when/how to use (e.g. to rebut a reputation-damaging claim that they said X). Claims that don't come accompanied with a specific attention timestamp (who said what, starting when and ending when, and perhaps where) will be presumed to lack credibility: and specific claims can then be refuted by the alleged actor. 

Those demanding credible information can then limit their sources to pre-AI-slop sources, + attention blockchain-authenticated expert sources. Individuals might become increasingly willing to connect to this incredibly privacy-invasive tracking in order to facilitate their work being included within the corpus of the credible.