An Under-Reported Serious Danger of AI: Destructive Political Ideologies Automated as “Fair Machine Learning”

And how I went from barely paying attention to politics for many years to suddenly devouring books and substacks on the topic

Introduction

The first academia dysfunction that started my path towards reinvigorating a thirst to understand the world around me occurred upon my being completely blindsided to the presence of what is more broadly defined as the “Social Justice”/”Diversity Equity Inclusion (DEI)”/”Woke” radical ideological force.

Here I present the specific history of my reawakening of intelligentsia class consciousness from encountering this. In particular, it applies in the context of what is called “Fair Machine Learning”, these are regulations on AI and AI algorithms that adjust the outcomes, so as to enforce “Fairness”. Yes, one of those Orwellian words that means “Goodness” used publicly, but some esoteric ideological social engineering privately.

When I had first joined the project, I naively thought the only Fair ML happening was simply the assistance in enforcement of Title VII and Civil Rights law – that, for instance, a highly qualified candidate isn’t turned away simply on account of their, e.g. race. I was told it was to assist in finding careless oversights that are inadvertently callous – for instance an automated system recommending candidates to hire who reside closer to a firm’s business location but an analysis finding that wealthier people are more likely to live in that neighborhood.

However, during the course the project, at one point I found out that an entirely different objective were the order of the day. Specifically, those enforcing Equity, formally Demographic Parity or Equal Impact, across different Demographic Groups.

It feels a bit odd writing this compared to my other articles and planned articles for the future. In particular, in my comprehensive investigation I saw no erudite defense of it, and intellectuals on the left are by and large completely embarrassed for their tribe because of it. So “dunking on Woke” is kind of a cheap shot, like beating up a much smaller opponent. It even feels like it is a sort of class assignment to be a proper substack intellectual, can you make a comprehensive argument for common sense.

Another way to look at it is that the problems with Equity based Fair ML are overdetermined, there are many various strong lines of reasoning and evidence that one can use to show that this would be an unmitigated disaster of a policy. It’s more about the choice of which one(s) to showcase.

But colleagues requested a summary and so here presents my attempt to present some, even though far from all, of the arguments to suggest that AI-automated DEI-ideological fairness would be nothing short of catastrophic as far as the set of highly likely adverse consequences.

History and Background

As noted, my understanding of the Overton window regarding fairness was very naive. Beyond the example of hiring recommendations based on proximity, there is also the possibility that even without a picture or name, a recruiter can identify cultural cues that suggest a race or ethnicity of the candidate, and thus it can make sense to have some means of identifying truly racist practitioners, that is resolutely not hiring a qualified candidate solely due to their gender or race. Even this should be treated delicately – indeed stereotypes being statistically true is one of the few well replicated findings in psychology. And so what if the recruiter is not identifying ethnicity but merely signs of a subcultural mileu that exhibits norms that are counter to productivity and reliability, but statistically correlates with some demographic identity?

Such were the technical questions I thought we would be wrestling with as part of the project. Now, I had heard about affirmative action in years past. But I genuinely thought it was something that was tried in the 1970s, turned out to be an unmitigated disaster, and discontinued. Then a couple black preacher-turned activists appeared in the 90s talking about it but they also talked about reparations and so people didn’t take them too seriously by and large. I had no idea that equality of outcomes thinking was present, let alone widespread enforced, beyond fringe Marxist social science faculty and perhaps some homogenously more left-wing enclaves like Portland Oregon or Norway. Maybe being born and living my early childhood during the course of the Fall of the Soviet Union gave me the End of History a la Fukuyama illusion that Bolshevik Communism and its related ideologies had been vanquished and comprehensively defeated.

Early on in the project, a paper was presented by the PI to the group that used optimal transport, the problem of moving one probability density mass to another, to adjust discrepancy in outcomes across demographic groups. I thought, naturally, that this was a technical exercise and no one with a sensible mind would implement such a thing. But I pondered how could this nice mathematics be actually useful? Then I brought up the extension and worked extensively with two other senior researchers, from top western institutions that are part of the consortium, to develop a scheme that was conditional on a set of Merit based features. That is, given the complete relevant set of Merit-based attributes, consider adjusting the probabilities of outcomes discrepancy. Developing some theoretical foundations and a specific algorithm, we suggested it to a PhD student on the project to work on implementing it in code. However, with no progress for some time, I was informed by the PI that the student had more important things to do first. Future inquiries led to the same responses, and a complete and expensive waste of many senior researchers’ time. Now, then I never thought it was a political thing, I had observed a pattern that the PI pursue his mathematical interests above what is useful for an applied project, so I figured something like this was going on.

There are various metric of fairness, each measuring a particular statistical relationship of outcome relative to the other features and outcomes across different demographic groups. Reading about the subject, I realized the significant component by which different political philosophies correspond to different notions of fairness. Having studied Philosophy, including Ethical and Political Philosophy in my youth recreationally but extensively, I decided to start a Philosophy paper aiming to provide a transparent taxonomy by which one can divide the various different fairness metrics across far left, center-left, center-right, and far-right perspectives. I identified center-right positions as interventions supportive of meritocracy and against ill will in the marketplace – someone refusing to hire people of a certain race regardless of qualifications being the primary target. The center-left would be more active in ensuring, for instance, a floor on statistical performance of an AI system across groups and avoiding unintentional discrimination such as the example of recommending candidates based on proximity to the work place, who ended up being more wealthy based on relative housing costs. Talking points for each would be grounded in standard Rawls-Nozick political center debate iteration. Then the far-right would decry any intervention in the marketplace and have no place for the whole enterprise. Then the far left is associated with Equity based fairness as defined by Demographic Parity and Equal Outcome metrics, as a complete wish to do away with market signals altogether, deprioritize merit, and have the outcomes socially engineered by the state.

Before a meeting for the entire consortium, I suggested to the PI I do a presentation on the ongoing paper and have an open discussion in the group and get a sense of the convictions of the participants and begin promoting open debate on efficacy for the project. The PI contemptuously refused. This and another matter made me immediately decide to leave the project. And my naivety about academia being a place of freedom of discussion and debate and comprehensive transparency of intentions was instantly shattered, and confirmed by subsequent research.

I looked at his publications on the topic, and they were heavily weighed towards Demographic Parity and Equal Impact!

I was shocked. Still naive, I thought this sort of policy wasn’t anything anyone believed in other than some blue haired UC Berkeley students and Claudine Gay. And Claudine Gay was discredited and so no one took DEI seriously. Here I was, shocked, stunned, completely disturbed.

I should have checked sooner as to the intentions of the project, and hopefully could have identified their nefariousness and left much sooner. But perhaps then I wouldn’t have had the impetus to devour information on history, economics, philosophy, sociology, psychology, etc. in an effort to find out WTF is going on, how did this happen? And perhaps then this substack wouldn’t have happened and more largely speaking, my class consciousness as a member of the intelligentsia not arisen.

After I left the project, relentless hostility from the PI ensued. The details of which I won’t go into.

Noting the presence of this Bolshevism having a habit of stifling free speech, I decided to bide my time, and wait until I received a permanent position by contract at my University before getting the truth out there. So here we are.

Fairness is an Ideological Subjective Value Narrative

The first section is demonstrating that Equity-based fairness is not aiming to fix any recognizable problem, that is, before going into the significant ills that are its consequences, here I will describe how 1) generally not what the general population wants, 2) does not correspond to any reasoned policy intervention, but is just ideology, and 3) tries to solve problems where none exist, except if you believe in a particularly radical ideology.

First, the notion has run into serious legal difficulties in the US, for affirmative action at Universities was declared illegal by the Supreme Court.

Ames v. Ohio looks to confirm that Equity across all markets actually violates Title VII Civil Right law, making DEI against federal law.

Polling suggests that it while people like the idea of promoting diversity in general, they oppose, by a significant margin, active discrimination for the benefit of minorities.

Moreover, it has failed in almost every state referendum that has been conducted.

On to the second point: you cannot find any technical analysis and description as far as why mandating quotas or equalizing outcomes in markets would yield a rise in social welfare of some population.

Let us trace back the intellectual Anthropology of this movement.

Many academics blame Critical Theory, as developed by the Frankfurt School and the French post-structuralists. However, I think that is a low resolution explanation. As argued in my work on the Epistemology of causality here, see the Social Sciences discussion, the problem is not the writings of the Frankfurt School per se, but the hubris of saying “complex description of comprehensive alienation” and “statistical disparities” implies any particular set of interventions. There is no argument by which post-structuralism or Frankfurt School alienation Hermeneutics suggests that widespread enforcement of Equity in market outcomes should be instituted and will fix these issues. I have an upcoming article in more detail on how I think the early Critical Theorists just played a troll game against everyone, simply pursuing their own personal advocacy from the status gained from their writings and the the right’s enduring stupidity problem.

Even blaming Marxist theory in itself is not quite apropos. Marx said “from each according to his ability, to each according to his need”, not “from each according to as much as it doesn’t make others feel bad about themselves”. For all its faults, the Soviet Union did, outside of Stalin era Lysenkoism, value scientific capability for rigor and reality-tested ingenuity. The physical and mathematical sciences and engineering were all world class behind the iron curtain. By contrast stands this insistence to prioritize not hiring “50 year old white men” by, for instance, the OceanGate CEO, in designing the Titanic tour submarine which disappeared in the deep ocean.

Ultimately, the ideology behind equity-oriented fairness is purely activist, just asserting that a certain state of affairs must hold, without any form of political economic analysis or even general principles that form some intellectual foundation.

The intelligentsia lineage of the ideology that culminated in advocacy for this policy was initiated by Paulo Freire, a Marxist sympathetic to Mao Zedong, who wrote about revolutionizing education. This, and his contemporary American protege, that is the intelligentsia milieu around Kimberlee Crenshaw, was written in the language game of political tactics, rather than political economic theory, a la What is to be Done written by Lenin before the Russian revolution. As a testament to their non-scientific form of pontificating: rather than publishing in well known respectable social science journals, much of this literature appears in their own venues like the Journal of Queer Theory, which have been demonstrated to be a complete and utter joke.

Of course, people of certain minority races have, on average, lower incomes and suffer from greater poverty, crime and other social ills. But where does this directly imply that the solution lies in equalizing the outcomes of all market competitions across these groups?

Rather, if we were to be truly epistemically virtuous, the appropriate subject matter for a specialized extension of Development Economics, that is the study of economic growth and development, specified to a community/group within a nation rather than considering an entire nation’s trajectory. Development Economics is a deep and nuanced subject, with a number of historical, cultural, institutional, international economic, technological, etc. factors having an influence of mixed indeterminable complexity on the process. Blind and blunt interventions are agreed to have a significant risk of doing more harm than good in this field.

Indeed, one can easily see that there are even government policies and institutional factors that cause racial disparities in life outcomes. For instance Savage Inequalities describe how the reliance of schools funding on local property taxes creates sharp differences in public schools. This policy together with cartel behavior on the part of the Teachers Union effectively shuts out establishing a real floor on school infrastructure funding and mechanisms by which talented students from poor socioeconomic zones can be discovered throughout merit testing and attend elite programs, i.e., the most robust means by which individuals from such communities can lift themselves out of poverty.

Finally, that is on to the third point: in some cases, in particular gender, no actual problem exists, that any fairness measures would seek to address, in the first place.

We observe a landmark study that showed that relative greater hiring for “male stereotype” jobs (construction, computer programming) for men has significantly decreased and now the relative acceptance of women for these jobs is even slightly greater than men, whereas relative greater hiring rates for “female stereotype” jobs (teaching, nursing) has remained stable at a significant level of advantage towards women.

When a request was made at Google to investigate pay equity, it found that it was systematically discriminating against men.

One of the common ideological myths is the so called “gender pay gap”: the observation that statistically, mean earnings for men are higher than women in all OECD countries.

We can observe, however, that when we split the data by married versus single individuals, young unmarried women are now making more income than young unmarried men. The earnings of married men are significantly higher than those of married women. Observing that 85% of consumer purchase decisions are made by women we can easily conclude that this statistical difference exists because of the sociocultural preference of a male breadwinner bringing income home for the woman to spend on the entire family.

So we have come to the place that clearly the “pay gap” is not any indication of “systemic discrimination” but simply the cultural choices surrounding family, and there is no reason to consider women to be a “protected attribute”.

We can observe that the progressive elite does have a particular ideological matter they pursue at all costs, however, in their assault on this cultural preferences. They make no secret of the fact that they seek policy interventions in order to socially engineer dual earner households with an equal distribution of time towards household tasks.

That’s not a matter important to me personally, I’m childfree so I, fortunately, have no choice to make in this cultural divide.

What I point out is that this is social engineering and ideology. Ideology that there are plenty of female intellectuals that vehemently oppose.

For an example of a psychiatrist who is a former feminist who couldn’t figure out a means of understanding her patients’ ills without questioning the 3^rd wave “equality” narrative.

Or, for the real lefties, using the frameworks of Lacan and Žižek to argue how this ideology is simply Capital taking authority over the potentially resisting family unit.

Or, if you want to check your Islamophobia, how about Muslim feminists decrying these principles as colonialism.

The most damning literature comes from Janice Fiamengo,

One interesting observation: you can see a clear correspondence to the oppressed vanguard faction bickering a la Animal Farm in the debates over TERFs (Trans-Exclusionary Radical Feminists). The virulent ruthlessness of some of the language these women use on each other is far harsher than anything chauvinist men are known to call women.

So we are indeed talking about ideological social engineering with far from universal support.

On the other hand, there are plenty of repeatedly replicated studies indicating significant biologically-enshrined differences between men and women as far as preferences and abilities.

Orwell mentions in Politics of the English Language that “equality” is a word that is often tactical – it means “goodness” when used publicly, but is some specific ideological intervention privately. We can consider that, for instance, it is at least to be considered an interesting definition of “gender equality” when the most “gender equal” country, Norway, accomplishes this by on average the government transferring $1.2 million more to each woman over a lifetime than she pays in tax, while the average man pays more in tax than he receives in benefits (there is surplus in total due to Norway’s government-owned oil industry).

Short Term Negative Consequences

Here I perform some basic Econ 101 reasoning to describe the disastrous economic consequences of enforcement of Equity based fairness in market outcomes and competitions. Through a number of mechanisms, overall economic output in the industries in which this will be applied will decline.

First of all, since merit is no longer the priority for reward, the aggregate performance of, e.g., the workers hired will be less than it otherwise would be. If group A candidates can produce $34, 35, 38, 28 value per hour and group B $18, 18, 30, 21, the optimal hire would be three of A and one of B, to get $34+35+38+30 = 137 total value. But if the hires must be equal, and merit is second, then one gets $35+38+30+21 = 124 total value. In general, however, the difference can be even greater than this, due to the Pareto curve, by which 20% of the workers perform 80% of the output, etc..In this case, e.g., if we consider these four to be the top candidates of a larger pool and change $34, 35, 38 to $94, 95, 98 then the diversity hire would be disadvantaged by a loss of revenue at $73.

While, due to blackout of any contrary positions in the literature, specific numbers are hard to come by. But as one example among a few I found, we have that with the introduction of gender quotas for corporate boards in Norway, “They find that firm value fell by more than 12 percent with every 10 percent increase of female board members” and that overall “the quota led to younger and less experienced boards, increases in leverage and acquisitions, and deterioration in operating performance, consistent with less capable boards.”

Beyond this, however, we need to look at the incentives these new price signals create.

If, for instance, incomes are equalized across groups and a quota system exists, then it would strongly incentivize small groups to form a cartel and shirk, or “soft quit”, perform in aggregate much less, knowing that due to regulations, the employer cannot do anything about the situation as far as their remuneration.

And in general, if incomes are equalized, then it would lower the incentive of working with greater productivity and effectiveness, because the employer would hesitate having to figuring out how to remunerate this person while maintaining the demographic statistical balance. In effect, if income is tied to the average, then it presents a commons problem with a game in which a race to the bottom is a Nash equilibrium.

In addition, now, considering the same example of equal wages across groups and quotas, we lose an important feature of prices, that is, wages, that they are meant to represent how much marginal productivity an individual could produce and correspondingly receive. With wages being variously perturbed from their market conditions, systemic shortages and surpluses will occur throughout labor markets.

The best workers, if tied back in their remuneration due to equalizing wages, would leave the sector of employment wherein this would take place, and the industry would see a flight at the top level of talent. They may start their own business, or move abroad, or pursue a different line of work that uses similar skills but isn’t subject to such regulations. Lately even starting your own business is fraught with DEI infestation, for instance NYC’s cannabis legalization store licensing laws resulting in 90%+ of the cannabis bought in NYC being from unregistered (formally illegal) dispensaries.

Finally, since the wage is relatively depressed compared to their marginal productivity, we would see white men drop out of the workplace altogether, as the relative value of leisure relative to work increases. This is what is happening statistically, as the labor force participation rate of white men has been in secular decline in the west.

In other words, if there’s no method by which working particularly hard for something would get him a good reward, then the choice an increasing number of white men naturally take is to just live a minimalist lifestyle and get by through collecting disability, driving a bit of uber, and slanging molly at the raves on the weekend to empowered boss babes.

Also for the people hired from the advantage brought by Equity enforcement, they would not be able to perform at par, on average, as the others in the team. This would create both circumstances of psychological guilt on the part of the hire, as well as encourage the stereotype of “oh, a diversity hire”. By contrast I know two brilliant black Physics friends in college and successful women Professors in my field, etc. whose strong work was enough to bring them success.

In many top ranked universities, they had to open remedial classes to accommodate a lower and lower level of understanding and standards. In the case of students admitted to University by reverse discrimination, they have significantly higher dropout rates than when they were selected by merit alone in the past. There are multiple studies demonstrating this empirically, here is one and two from the University of California

All of these factors have negative economic consequences. Work will be performed less well, both directly, and indirectly through incentivizing shirking and disincentivizing productive work relative to alternatives. The best performing talent will flee from these industries. In aggregate, all of these factors will cause a significant drop in the economic performance of the industry on which Equity-oriented fairness is applied, and ultimately lower the average wage across all workers in the industry, and in aggregate reduce economic output of the economy.

For some deeper analysis of economic institutional consequences of dropping merit as priority and the corrosive incentives of Equity ideology, see, for instance Michael Magoon’s writing here and here.

And for a compendium of references exhibiting unintended negative consequences, the failure to establish enduring diversity, and other evidence in the scientific literature regarding DEI’s negative consequences, see the diligently curated post from Lee Jussim here.

Medium Term Negative Consequences

Recall one of the most significant economic phenomena in recent times: deindustrialization, and thus the loss of working class manufacturing jobs, together with a stagnation in median real wages for the last 50 years while income to the top 20% and especially the top 1% and top 0.1% have sharply increased. Meanwhile, productivity of labor still increased in this period of time, which is anomalous to the stagnation in wages, i.e., in a free market, this should not happen.

Indeed DEI was one of the significant components of regulations requiring administrative and legal departments better suited for larger corporations, leading to an increase in the concentration, i.e., monopolization of the economy. The addition of bureacratic bloat at both the federal and firm level created what’s called the Baumol effect, whereby lower marginal productivity in another part of the firm’s operation, such as whole departments that add no benefit to the production of goods or services to the firm, depresses wages across all worker functions in the firm.

With more automated DEI all of these phenomena would accelerate.

Let us review some of the consequences of this. For one, we can identify that the stagnation in real working class wages has resulted in a significant turn to the negative in white middle aged men’s mental health, with, due to deaths of despair (suicide, drug overdose), are such as to cause 100k early deaths over 15 years .

There are additional follow-up effects. Indeed, this writer identified that the cause of the teen mental health crisis is less smart phones and more their parents ill mental health and intensive parenting due to competitive precarious economic conditions. In addition, men who lose their jobs tend to get divorced more, and in general men who make less than their wives get divorced more, and the increase in children without married parents has contributed to the increase in mental health conditions among young people, or Gen Z, morbidly speaking, there has been a notable spike in the suicide prevalence among the younger generation in recent years.

Next are the political effects.

Just like Marxism, a tangled web of signifiers light years from any reality on the ground, so is Equity-oriented Fair Machine Learning. Indeed all the words themselves carry with it pleasant pro-social imagery, while ultimately enforcing a particular tribe’s ideology. If the outcome are catastrophic, researchers are far removed from any accusations of culpability. As a general rule “In the absence of reality-testing, bullshit wins.”

At the same time, the associated ideological tribe is unequivocally aggressive and opportunistic at yielding its agenda. Indeed, a recent study, unreported by the media, observed significant spikes in antagonism, hostility, and a generally oppressive environment in workplaces that introduced DEI policy. If a policy was genuinely and clearly beneficial to society, it wouldn’t have been necessary to exert aggressive control of the narrative and severe hostility to those who spoke against it. With the savagery with which they pursued their agendas, the silent majority classically left-liberal faculty held their tongues. The confidence with which these ideologues pursue their agenda is a clear demonstration of the Dunning-Krueger effect, as it has been found radicals have less political knowledge than moderates.

As I argue throughout the blog, the rise of the populist right is because of the dysfunctions in the intelligentsia, if you are upset at the populist New Right, you should really take issue with what caused them to rise in the first place. For instance, if you want to hear some academics shit on the Orange Man and then talk about how him getting rid of the omnipresent DEI would be a big win for academic freedom, see here.

This advocacy also crowds out advocacy for the interests of the poor and working class. Gary from Gary’s Economics worked for a progressive think tank, wanting to talk about how, from his experience banking, he could write about how to best reduce inequality. But they were having none of it and focused on social justice and climate activism. So he eventually left in frustration and started his youtube channel.

And so, the Intelligentsia lost the trust of the general population.

Since left no longer represents interests of working class white men, they have no choice but go to populist right.

Consider 1) that 94% of jobs gained surveyed in the year 2021 went to non-whites, and similar numbers follow, suggesting DEI has fully degenerated into an Animal Farm “all animals are equal, but some animals are more equal than others”, then 2) the media says the economy is doing great, as per capita incomes are up - but median wages priced to a working class family basket of goods is down, and stock prices are up - while most working class people don’t own stocks, then you have a situation where the Democrats were gaslighting the working class in 2024, and if you’re in the intelligentsia and you are surprised by what happened, you are ignorant. And if you directly or indirectly contribute to the expansion of DEI in Europe, then you will see the bed you shall lay on, as I describe below.

Elections have consequences. For instance, it is because US AID was rife with funding DEI activism across the world, a lovely exercise in cultural colonial imperialism, as well as nice commissar kickbacks ($5 million for tourism promotion in Egypt?), that the population is agreeable to its financial gutting. Yet, there were also legitimate foreign aid. And so all those starving children in the Global South who lose out, that again can be fault laid at the hands of this ideological tribe.

It would be naive to not consider current events foreshadowing a European Parliament, as well as the nations of the EU, led by the AfD and National Front, etc. cutting European aid, if the EU does not stop its course of pushing Equity-oriented fairness and enforcing it with AI. It’s clearly on its way, with Orban, Meloni, the Swedish Democrats, etc. expected largely to retain power.

Long Term Negative Consequences

The Long Term Negative Consequences of Equity-oriented Fair ML is a matter of simple deduction.

Consider these readily agreed upon basic ground facts:

1) The political cosmology of this ideology states that the presence of any statistical difference in the outcome of a desirable outcome, e.g. salary level or positions of status and authority, that favor one group over another then this is due to systemic oppression of the first over the second group, and must be corrected for.

2) There is a general emphasis in the AI community for AGI, and being able to generalize better to new instances by learning the embedded underlying structure of principles in the operation.

3) We expect ever more resources to be devoted to AI making its actions increasingly more powerful, even if blunt, in grounded impact.

From these we can conclude, as basic Systems thinking, that such a Fair ML system, as it becomes increasingly developed, will attempt to generalize the principles, that is, the political cosmology of the intervention.

As such, it will perform inference on various data that it can find on the general principle of “any statistical difference in outcome is systemic oppression and must be corrected”.

Let’s consider what sort of empirical data is out there that is relevant.

Actually, there is an interesting data point from which to make this inference. There is indeed a demographic group that has a higher income, more positions of status in politics, entertainment and the sciences, and outshines the rest of the population on a range of life outcomes.

Can you guess who it is?

….

any guess?

…

…

It’s the Jews!

And so, we can conclude that, together with the priors:

1) there is increasing polarization and political radicalism in the present day both towards the far left and far right

2) people have a long history of using “systemic oppression” narratives involving the Jews as an excuse to persecute them

3) the current tension of public opinion on Israel regarding Netanyahu’s behavior. The social acceptance of the phrase “from the river to the sea” as acceptable language in some public environments.

Let’s not be naive, do you really think we’re that far from more development in equity-oriented fair machine learning facilitating something absolutely horrifying from the inference, on the data of their relatively extraordinary outcomes as a group, “...therefore the Jews are oppressing the rest of us and this must be corrected”?

Reflections

From the arguments above, let us now make estimates as to the consequences, as far as body count, of an implementation of automated AI Fairness, especially enshrined by metrics of Demographic Parity and Equal Impact playing a significant contributing factor in the causal dynamics. We can use the quantities cited above regarding deaths of despair and make additional extrapolations from additional economic and political consequences described above, and use previous historical instances to extrapolate mortality figures for the Long Term Risk applied to demographics in the present day. My estimate is:

1) At its current level of implementation in the EU AI Act, with the algorithms of this and similar projects’ algorithms, about a 95% chance of at least one hundred thousand early deaths in the next fifteen years.

2) With the likely continued pace of development of the algorithms, with considerations of the Short and Medium term consequences, a 70% chance of causing at least one million early deaths in the next fifteen years.

3) With the likely continued pace of development of the algorithms, with a 30% estimate of the development described in the Long Term Consequences, that would yield at least 10 million early deaths in the next fifteen years.

While one may debate the precise probabilistic forecasts, the underlying qualitative facts are undeniable: 1) AI-automated Equity Fairness WILL significantly increase the mortality rate in Europe, 2) Moreover, there is a nontrivial chance that it will facilitate mass state-sponsored violence.

Considering then how obvious and overdetermined these stark realities are, since coming to face this nefarious research program, I’ve been very interested as to how the hell this could have happened. How it was that this ideology infected the intelligentsia and have thoroughly studied its development as far as an incentive for signifiers that yield catastrophe on the ground. The study of the Symbolic Capitalist class (academia, media, PR, etc.) by sociologist Musa Al-Gharbi has some very insightful work, such as the article subtitled: Professionals often define “Racist” in ways that serve their own interests – to the exclusion (or at the expense of) the preferences and priorities of the genuinely marginalized and disadvantaged, as well as a book titled We have Never Been Woke that describes the economic, sociological and political history of the Symbolic Capital Class.

And it’s real, the EU AI Act has passed Parliament and is on its way to being implemented. The same content coming to America at the state level.

As someone who is psychologically particularly prone to feelings of guilt and shame...to the extent that I have contributed to the project’s proposal writing, those are gallons and gallons of blood on my hands. My actions WILL cause mass deprivation, suffering and early agonizing death on a mass scale.

I experienced regular nightmares, a recurring one was seeing large piles of corpses all around me and feeling horrified while the PI was gleefully skipping on top of the dead bodies “what’s wrong with you? Why don’t you like fairness?”

The fact that I was truly bait and switched as far the as the real goals of the project was well ingrained in me in therapeutic discussions, and, of course, finally getting the truth out there as far as this blog article helps.

Otherwise, let’s end this article on a philosophical note.

One natural response is complete Nihilism. The fact that I contributed to Equality-focused fairness means my karma in this world is such that, no matter what I do in this life, the balance sheet is very negative, akin to that of serial killer or Congolese warlord. In that case, why bother ever considering morality or alleviating human suffering in any way? With that ledger stuck at negative why not treat life as purely for my own entertainment and pleasure? If I’m already partially responsible for mass suffering and death, declare a broad “Fuck It!”

There’s another perspective that I adapt from a hypothetical similar discussion from Mike Israetel: consider that you are a pilot and due to poor flying decisions, while at the same time not foolhardy enough so as to warrant legal criminal negligence, you had to eject out of a falling plane that ended up dropping onto a nursery and killing a school of children. Mike mentions that one can consider this as simply a done and done balance sheet, or, instead one can use that guilt to bring passion into developing some charity or other effort to improve the lives of impoverished children throughout the Third World, with a firm dedication towards alleviating their suffering.

Similarly, I can make a dedicated effort towards pushing for ideals of research that represent true epistemic virtue as well as the prioritization of making research results genuinely useful for humanity, in alleviating poverty, improving the conditions of the working class, and contributing to economic development in the Global South.

Let me now share some family history. My great grandfather had a large family and built a big house in the 1930s for them. However, the Soviet government decided that this relatively large house compared to others in the town violated Equity Fairness and sent him to the gulag and took the house over and made it into a Community Center (place where Party members got drunk). When Operation Barbarossa started the government sent him to Minsk as cannon fodder and that was what he ended up as. My great grandmother and grandmother and grand uncles and aunts survived by foraging the forest for some time.

And now we have yet again, my blood in one corner of the ring, and Bolshevism on the opposite corner.

The bell rings

Round two