I recently wrote “The Eucharist, Redux #6: AI Apologist” which highlighted the pitfalls of AI when it comes to apologetics. But my warnings…
AI is often quite wrong and will contradict itself.
AI is only as good as the data it is trained on.
AI can be coaxed into admitting its errors, but it’s still going to make them.
AI may well tell you what you want to hear, rather than what is right.
…apply to any topic. When it comes to programming questions, StackOverflow has been the go to source for programmers for a long time. It contains answers to most common programming problems written by actual programmers. It typically includes multiple solutions to any given problem as well.
The problem is that AI is trained on data sets like this. AI is only as good at programming as the material it has, and Stack Overflow is a huge contributor. As people start using AI instead of human-generated sources of information, the amount of novel solutions will dwindle as a circular system of regurgitation develops where AI trains on AI.
This is obviously bad.
Bruce Charlton has long noted an analogous process at work among humanity. In ages past, expert human knowledge was transferred from person to person through systems of guilds and apprenticeship. In general, those who possessed great skills were able to directly transfer those skills with minimal losses and often some gains.
When bureaucratic processes started to dominate society, information was transferred to written form and standardized. Instead of direct a person-to-person transfer of knowledge from master to apprentice, education became the norm. These days few teachers are masters in the subject that they teach, and textbooks are designed to avoid specialized knowledge in favor of the most average or approachable knowledge is.
This is how, despite the rise in apparent IQ,[1] society has been losing expert information at a relatively rapid pace, especially compared to the relative rise in population.
Take civil engineering and architecture for example. Designs for buildings and structures is better and safer than ever, but so much relies on software rather than expert knowledge. Just like calculators have taken away the need for people to basic arithmetic skills in order to function, so too has reliance on automated tools eliminated the need for every engineer to possess all the required information to make sound designs. Design defects are simply not as big a concern when the computer can automatically avoid or resolve them. So, even though the end product has gotten better over time, the level of expertise in the field has gone down.
AI is taking this to the next level. For a while the end product will get better—as we are still in the early days of AI development—but at the cost of an expected massive decrease in the total amount of human expertise. AI will be able to offset this somewhat by ingesting new human-produced outputs, but this will only slow the apocalypse, not prevent it.
And things will be more average, more mundane, more boring. True novelty and innovation will be hard to come by.
Footnotes
[1] See Bruce Charlton’s clarification in this comment below.
@Derek – You are far too all-round optimistic IMO!
History is linear, not cyclical, and there has never been a time as degraded and corrupted in its motives as now, in The West. This *will* ensure that current so-called AI will Not improve anything that is functionally significant, but will instead make matters worse on average and increasingly.
This should really be obvious by considering the top-down Western Establishment provenance of current “AI” (including the dishonest naming!) – and the coordinated multinational political-bureaucratic-media campaign by which it has been imposed.
Point of information – you mention the rise in “apparent IQ”. IQ needs disambiguation. Strictly, IQ refers to test results, and the biological intelligence “g” is what IQ testing measures – albeit in a comparative fashion
(i.e. IQ testing measures whether one person or group is higher, the same, or lower in g than a comparative person or group – IQ testing does not quantify intelligence, nor does it relate the IQ number to an objective measurement.)
Raw Scores on IQ testing increased for several decades in developed countries. (i.e. the “Flynn” effect – or more accurately, as Flynn himself acknowledged, the “Lynn-Flynn” effect – since Richard Lynn first discovered and described it and published the result in Nature. Flynn then did extensive work to confirm and measure the effect.)
But rising IQ test scores is of no more fundamental significance to measuring long term trends in underlying human intelligence (“g”) than is the inflation of other kinds of examination results.
After all, IQ tests were designed to be comparative in cross-sectional tests of populations – is was designed to put IQ testees into rank order. That rank order of IQ scores in a population is (on average) maintained over time – but ranking does not reflect the underlying biological intelligence of that population, which in fact very-probably declines from early 20s age (in men, earlier in women – i.e. objective measures of reaction time and brain processing speed get slower through adult life, accelerating into old age).
I think Michael Woodley and others have now confirmed what he and I first asserted with evidence from objective reaction times data, back on the afternoon of 28 February 2012 (!) – that real, biological, underlying average intelligence has been declining in the UK and other developed nations for (probably ) 200-plus years:
https://charltonteaching.blogspot.com/2012/02/convincing-objective-and-direct.html
http://iqpersonalitygenius.blogspot.com/2017/03/5-years-since-woodley-and-i-announced.html
Current AI is overwhelmingly a demonic instrument in the spiritual war of this world. That is its core significance, swamping all other particular applications.
Bruce,
This..
…real, biological, underlying average intelligence has been declining in the UK and other developed nations for (probably ) 200-plus years…
…is the reason I said “the apparent rise in IQ.” I’m well aware of all the things you have described here, having written quite a few articles that discuss these things. I suspect you didn’t read my site when I wrote them. Just see how many times I’ve quoted from Genius Famine here and (especially) on Twitter (or my citations of Woodley)!
You are far too all-round optimistic IMO! … Current AI is overwhelmingly a demonic instrument in the spiritual war of this world. That is its core significance, swamping all other particular applications.
I’m a programmer and mathematician, so I don’t ascribe mystical attributes to AI. The biggest threat from AI has always been the people making it and using it, not the technology itself. Blaming AI for evil is like blaming a gun for gun violence. AI is just another tool in the toolbox, even if its adoption is like replacing a hammer with a nuclear bomb.
AI is a glorified search engine. People forget that before the popularity of search engines, people found other sites by going to sites they knew about and selecting things from their lists of curated sites of interest. Although blogrolls still exist, most of this went away when we gave up individually curating articles and sites in favor of Google telling us what to read. It’s analogous to the master-apprentice vs. bureaucracy distinction I made in the OP.
Google search is just as evil as AI. It is equally a demonic instrument. But we are used to it, we have accepted it, so we don’t think of it in those terms.
AI is going to become even more pervasive, but it’s not going to alter the default state of the heart of mankind, which is already fallen. It’s going to revolutionize society like the internet and iPhone did. But very little in the heart of man will actually change. People already willingly believe almost any and every lie, and are willing to abandon any and every virtue on a whim for its inversion. AI will simply continue that trend, just in a new and different form.
I’m optimistic because I’ve long since given up the idea that it is the job of Christians to fix the world. Even if Christians go back to the persecution and suffering described in 1 Peter, I will still remain optimistic. I have hope in what is to come. I encourage my fellow brothers not to despair.
The 2020 lockdowns were undeniably demonic, but they also clarified what was and wasn’t the church of Christ. Each church and denomination declared its ultimate allegiance to either Christ or the State, and very few have publicly changed their allegiance since then. We have moved into an era where right and wrong have become increasingly obvious to distinguish. AI will make this even clearer.
The biggest concern won’t be that AI is evil, per se, rather it will be that the vast majority of people have evil in their hearts and have actively chosen it. No man will be able to say “I didn’t know!”
It is like your talk of Litmus tests.
Peace,
DR
Bruce,
I’ve added a footnote to your comment in the OP.
Peace,
DR
When bureaucratic processes started to dominate society, information was transferred to written form and standardized. Instead of direct a person-to-person transfer of knowledge from master to apprentice, education became the norm. These days few teachers are masters in the subject that they teach, and textbooks are designed to avoid specialized knowledge in favor of the most average or approachable knowledge is.
YEP!
The long-predicted by the MGTOWosphere and then the Roissyosphere(which through its blank slate loving offshoot the manosphere was supposed to prevent by ”easily” teachable by game ”masters” like Jack to all the novice peons pout here w/o a 140 I.Q.,
to MEN to be wanted and attractive through game)scapegoating of unwanted, unattractive MEN(associates them with Terrorism for simply existing) has also begun:
https://www.youtube.com/watch?v=PzB-GQqrE7w
Starmer: Britain faces a threat new of terrorism from “loners, misfits, young men in their bedrooms”
The News Movement
145K subscribers
Subscribe
164 views Jan 21, 2025
Britain faces a new threat of terrorism from “loners, misfits, young men in their bedrooms” following the Southport murders, the UK Prime Minister has said.
Axel Rudakubana, 18, pleaded guilty on Monday to murdering three girls at a Taylor Swift-themed dance class in Southport, Merseyside, in July.
Despite contact with state agencies such as Prevent, aimed at countering terrorism, authorities failed to stop the attack which claimed the lives of Alice da Silva Aguiar, nine, Bebe King, six, and Elsie Dot Stancombe, seven.
Addressing an inquiry launched on Monday, Sir Keir said failures of state institutions in Rudakubana’s case “frankly leap off the page”.
Mr Rudakubana, from Lancashire, will be sentenced on Thursday, he pleaded guilty. He is not expected to receive a whole life order because he was 17 at the time of the murders.
Unrest erupted across the UK in the wake of the Southport attack, with mosques and hotels used for asylum seekers among the locations targeted. More than 1,000 arrests linked to the disorder have been made since the attack, and hundreds have been charged and jailed.
The ”manosphere” kept claiming this was going to happen since ell!ot r@dgers in 2014 did his stuff, yet their way of combating it was to needlessly make everything overly complex with new terms, charts, graphs, and the pushing/sponsoring of George Soros-paid leftist ”Conservative” Republicans as supposed non-deep state soros paid leftists.
Good job, supposedly non-Soros-paid guys!