AI's Insurmountable Flaw: "Mass Regurgitation of Misinformation"

These immense hidden costs will not show up in GDP until they collapse the entire house of speculative gambling cards propping up the global economy.

I approach all AI topics with several things in mind. One is the nature of problems, which implicitly define what qualifies as solutions, and the resulting incentive to define the "problem" such that the "solution" happens to be the one we own and control.

So the "problem" AI solves is "corporate profits are too low," and so the "solution" is to replace costly human labor (made costlier by SickCare insurance and taxes on labor) with "cheaper" AI (cheaper because the full costs are hidden or subsidized).

My other lens: the economic, social and cultural consequences of AI as it is and AI hype, a topic I've explored most recently in Is AI Reversing Anti-Progress or Is It Accelerating It?, AI Data Centers Are Not the Railroads of Today and Inequality, AI and Digital Life Are Undermining Society.

Correspondent Mike Fasano recently submitted a succinct and telling summary of AI's insurmountable structural flaw: AI's inability to discern the difference between truth and falsehood, be it intentional misdirection / misinformation or errors generated by AI hallucinations, a systemic flaw which he summarized as mass regurgitation of misinformation:

*           *           *

"I read you post on AI and railroads. Here is another observation.

So far, AI has only regurgitative intelligence. It--at best--can collate and respond to queries on masses of acquired data.

But what if that data is wrong?

Who now believes the inflation or unemployment statistics? Virtually every human knows that those statistics are false.

Does AI know that?

And the problem goes much deeper.

The former editor of the New England Journal of Medicine, Marcia Angell, noted:

'It is simply no longer possible to believe much of the clinical research that is published, or to rely on the judgment of trusted physicians or authoritative medical guidelines. I take no pleasure in this conclusion, which I reached slowly and reluctantly over my two decades as an editor of the New England Journal of Medicine.'

That being the case, can we rely up AI medical advice?

And that problem goes beyond medicine. It is now generally conceded that the inability to replicate scientific studies of any type has give rise to a 'replicability crisis' in science. Can we trust 'science' that cannot be proven to be accurate?

Any adult past the age of 40 knows that the above listing of questionable information sources is just the tip of the iceberg. We live in a sea of 'official' but false data.

Railroads could transport grain to cities, minerals to factories, manufactured goods to those needing those goods. That served a public purpose.

But what is the use of the mass regurgitation of misinformation? And is anyone subtracting the losses engendered by the utilization of inaccurate information from GDP?"

*           *           *

Thank you, Mike, for clarifying an essential point: the foundation of all "value" is fact, truth, accuracy and the transparency, replicability and accountability of the processes validating fact, truth, accuracy. If AI is incapable by its nature of validating all these, it's worse than useless--it's destructive on a system-wide scale.

The evidence of the systemic destruction is already overwhelming. Bogus "scientific papers" are already proliferating at an accelerating rate, making the task of identifying incorrect and fabricated (i.e. hallucinated by AI) data, processes and conclusions impossible due to the scale of the misinformation and the difficulty of identifying the misinformation buried inside superficially legitimate papers.

With both scientific and economic data and analysis now untrustworthy without exceedingly expensive, time-consuming vetting by human experts, where does this leave the "AI will automatically generate superabundance" hype? What's already clear--but inconvenient--is the mass adoption of inherently flawed AI is undermining the foundations of "value," however we wish to define it.

And as Mike also points out, this undermining of value has a financial consequence. We all know Gross Domestic Product (GDP) is a superficial, distorting measure of "prosperity," and the structural distortions of GDP (Waste Is Growth) are amplified by the hidden destruction of transparency, replicability and accountability by AI slop, whether intentional (malicious, deceptive, fraudulent) or as the unavoidable consequence of AI's core flaw.

These immense hidden costs will not show up in GDP until they collapse the entire house of speculative gambling cards propping up the global economy. Only then will the structural damage being wrought by our increasing reliance on tools that cannot discern the difference between fact and fantasy / fabrication / hallucination become visible.



And by then, of course, the damage will be irreversible without extraordinary costs and sacrifices, sacrifices few will volunteer to bear.

Remember that AI isn't "thinking," "understanding" or "making judgments": AI tools are engines of linguistic automation, not engines of understanding. The simulation is not the thing simulated. AI is not a "mind," it is a prompt and a probability distribution.

AI and human intelligence are drastically different--here's how.




My book Investing In Revolution is available at a 10% discount ($18 for the paperback, $24 for the hardcover and $8.95 for the ebook edition). Introduction (free)


Become a $3/month patron of my work via patreon.com

Subscribe to my Substack for free

---

NOTE: Contributions/subscriptions are acknowledged in the order received. Your name and email remain confidential and will not be given to any other individual, company or agency.

Thank you, David M. ($75), for your marvelously generous subscription to this site -- I am greatly honored by your support and readership.		Thank you, Julie W. ($3/month), for your wondrously generous subscription to this site -- I am greatly honored by your support and readership.