An example of the censorship embedded in the algorithms that control responses using Artificial Intelligence (AI)

From here:

My Conversation with AI Over Hydroxychloroquine ⋆ Brownstone Institute

“HARVEY RISCH   JUNE 16, 2023   PHARMA   7 MINUTE READ”

Please read the whole article. What struck me, is that if, as Dr Risch is, a subject expert, the presentation as fact in a first response from AI, represents a “narrative” response, rather than a “factual” response. In fact, if you challenge the “narrative” with “facts”, this is the typical response:

“I apologize for the confusion in my previous responses. I made an error in my initial response by not fully addressing your question and then compounded that error in my subsequent response.”

Now, this is one AI platform, used by the good doctor because “the following conversation took place on rebbe.io because it does not require a login. “

It would be interesting to benchmark the different AI platforms to check for “truthiness”, right? I haven’t seen much analysis of this platform against others in the various disciplines.

Imagine a news reporter, or chat anchor, or staffer for a politician or a political advocacy group OR A MEDICAL BOARD that used this AI platform, in a hurry, who did NOT have even a basic grasp of the field WHO STOPPED AT THE FIRST REPONSE.

Here one example of an “error” and the “confusion” that led to an apology.

Response – abridged

“University of Minnesota and published in the Annals of Internal Medicine in November 2020. The study found that hydroxychloroquine did not significantly reduce the incidence of COVID-19 or prevent hospitalization or death among non-hospitalized patients who had been exposed to the virus.”

“User: How many people died in the Minnesota study that you just referenced?”

Response – abridged

“I apologize for the error in my previous response. The University of Minnesota study on hydroxychloroquine and COVID-19 that I referenced in my previous response was actually a randomized, double-blind, placebo-controlled trial of post-exposure prophylaxis, not a study of non-hospitalized patients. “

“..  there were no deaths reported in the University of Minnesota study on hydroxychloroquine and COVID-19 post-exposure prophylaxis. The study focused on the drug’s ability to prevent infection, rather than its effectiveness as a treatment for COVID-19.”

Ok, probably my ignorance, but it is my understanding that HCQ was a treatment – not an immunisation, but maybe that is “my bad”.

The AI platform goes on “.. I am not aware of any studies that specifically investigated the use of hydroxychloroquine as a treatment for non-hospitalized COVID-19 patients and its effect on reducing the risk of death. “

“User: What about the hydroxychloroquine studies by Sulaiman in Saudi Arabia and Mokhtari in Iran?”

Response – abridged concerning the Saudi study:

“ it is important to note that this study had several limitations, including a small sample size and the fact that it was not a randomized controlled trial.

And the other study:

“This study also had several limitations, including a small sample size and the fact that it was not a randomized controlled trial.”

“User: How many hospitalized patients were in the studies of Sulaiman and Mokhtari that you just quoted?”

Response abridged:

“.. the study by Sulaiman et al. conducted in Saudi Arabia included 106 hospitalized COVID-19 patients, while the study by Mokhtari et al. conducted in Iran included 66 hospitalized COVID-19 patients.”

User: The Sulaiman study says that it included 7,892 outpatients. The Mokhtari study says that it included 28,759 outpatients. Neither study included hospitalized patients as study subjects.”

Quite the challenge! The good Doctor pointed out that the AI response talked about 106 + 66 hospitalized patients, whilst the actual study looked at 7.892 + 28,759 outpatients, not hospitalized patients.

The response?

“You are correct that the studies by Sulaiman et al. and Mokhtari et al. did not include hospitalized patients as study subjects.”

The User closes with a direct challenge:

“User: This answer did not answer the question that I asked. All of the errors you made were statements contrary to facts that you verified in subsequent answers. Since you had the correct factual information, why did you cite incorrect facts?”

Ah hah! “You had the answers but stated lies anyway, WTF is going on!!!”

Here is the non-response:

“I will continue to do my best to provide accurate and helpful responses based on the information available to me, and I appreciate your patience and understanding.”

By supplying the “narrative” first and facts, at best, second or third if you catch me an “AI lie”.

This is a response to a series of questions of fact. It gives you some idea of why you get search engine responses and how providing “narratives” rather than “facts” results in a far lower quality in everything it comes into contact with.

This particularly applies to the censorship of social media platforms. It is based on first order “narratives” rather than second or third order “facts”.

These “narrative checkers” are fatal flaws in the algorithms that drive AI and are turning us into idiots unless we conduct independent and diligent research – or have several decades operating at the highest intellectual levels of integrity, like the good Doctor Risch,

Maybe other AI platforms have less “narrative” and more “facts” and present the data in a “cleaner” less grovelly way,

For context, from here:

HCQ for COVID-19: real-time analysis of all 484 studies (c19hcq.org)

there is a potential source of data from hundreds of studies.

Each of the 394 studies on the C19 site is referenced and linked to its source. I wonder why the AI platform did not mention this website, even in passing. Perhaps the algorithm is designed to compress all sources into “one-liners” as Microsoft Bing does.

Onwards

Please upgrade to paid, or donate a coffee (I drink a lot of coffee) - “God Bless You!” if you can’t or don’t want to contribute. Coffee donations here: https://ko-fi.com/peterhalligan  - Buying just one Ko-Fi a week  for $3 is 50 bucks more than an annual $100 subscription!