AI and Humanism

Balázs Kis
May 27
7 min read

Updated: May 28

I'm brave enough to assume that every member of the AI Localization Think Tank is a humanist. We never talked about it explicitly, but this is what my experience tells me. By 'humanist', I mean the following with respect to AI:

All technology development, including AI, must work for the clear benefit of human beings collectively as communities as well as individual humans. Technology, including AI, must provide these benefits without causing harm to said humans (communities or individuals), or collateral damage to other communities, individuals, or the natural or artificial environment.* [1] [2] [3]

You might believe this humanist approach is a truth universally acknowledged among researchers and developers of AI. But it is less common, or popular, than you would think. The harms caused by public AI services suggest that their owners or operators think differently, and do not always act with humans in focus. Or, they do not think: they do not have any ideological or ethical foundation other than their own financial bottom line.

Here at the Think Tank, we keep dissecting the various behaviours of AI models, including their sycophancy, hallucinations, and nonchalance about the mistakes they make.

In this article, I want to do something else. First, I’d like to highlight some anti-humanist ideologies that might have informed AI developments thus far. Second, I’d like to call attention to two humanist projects about technology that deserve more of a spotlight than they currently receive: the Center for Humane Technology and the Digital Humanism project at the Technical University of Vienna.

Post-humanists, trans-humanists, and the rest of us

Humanism faced (and to be honest, partially deserved) criticism and challenges from two main directions:

Post-humanism: Human-centred approaches to life and thinking were labelled as ‘arrogant’ because they seemed to refer back to the biblical “promise” where God, in the book of Genesis, puts all of nature under human rule. This and the evidence of human-inflicted climate change, and all the other harm done to the environment, led to the idea that humans, and humanity as a whole, are an altogether negative force in the universe. According to post-humanist thinking, humankind (the Anthropocene) will end, and it will be a welcome development for the Earth and the universe. [4] [5]
Trans-humanism: Trans-humanists think that the achievements, the wealth, and power of humanity must be preserved for eternity, or for as long as physically possible. This wealth and power must also be owned and controlled by the descendants or successors of humanity. This view serves the heirs of humanity in the distant future and does not have any regard for individuals or organisms currently alive. Trans-humanism acknowledges that the human body might need to change or evolve, and often proposes to accelerate this change, for example, by making AI devices where human minds and consciousness can be “uploaded”. This view also assumes that only a select few humans are “enlightened” enough to know how this can be achieved, and they are entitled to control the rest of society. [6] [7] [8]

On the one hand, post-humanism isn’t interested in technology development, and individuals representing post-humanist ideals do not care about creating AI, humane or otherwise.

On the other hand, the trans-humanist worldview can be recognized in the wish for a “techno-authoritarian” society, as well as in the characteristics of social media. Techno-authoritarianism emerges when society allows the creation of a few technology corporations with enough resources and power to effectively control entire communities or countries. [9] [10] [11]

The techno-authoritarian worldview has no issue with replacing human work with automation, since it is interested in the preservation of achievements, wealth, and control. There is a popular misconception that companies that replace human work with AI also damage their own markets. No: it is possible to create an economy where buy and sell decisions are both made by automated agents, without the need for humans to be alive to oversee these. There are abundant examples in the world of the stock exchange or crypto, or more recently, travel arrangements.

To be fair, not everyone jumps on the AI bandwagon with a clear, sinister trans-humanist plan. Many companies mindlessly replace human work with AI because they manage by cost-cutting;, and there is a popular (albeit unproven) belief that AI costs less in most, if not all, work environments. [12] [13]

A return to human(ism)

Back to humanism: as a humanist, I believe that existing life has value in and of itself. I believe that every human life is immensely, unfathomably valuable because all humans have the potential to create more value, to cater and care for life, and contribute to the “achievements” of humankind. This is not transactional: your value does not depend on what you contribute. Your life is valuable because you, like all humans, have the potential to contribute.

I also believe that it is possible to build technology, including AI, that respects and benefits human communities and individuals. Some companies claim to do so. I make no assumptions about the motives of AI and technology providers, although some of them have made their positions abundantly clear. In some cases, companies claim to respect humanist values, but there is a discrepancy between this and the behaviour of their products. I would use the term “humanism-washing” to describe this, adding to the growing list of “washings” that refer to performative approaches to ethics.

I think it benefits a company to clarify, to themselves and to the rest of the world, where they stand with respect to fellow humans. In my experience, we think too little about our values and motivation.

Two initiatives caught my eye much earlier than November 2022. They explore ways to create “humane” technologies because they recognize the addictive and exploitative tendencies of modern technology development. This escalated when social media became commonly used and was made into a tool for manipulating and distorting people’s world views.

The Center for Humane Technology is a think tank founded in 2018 by Tristan Harris, Aza Raskin, and Randima Fernando. In the 2010s, Harris worked as a design ethicist at Google, and later began to speak out against technology that creates addiction and consumes most of the users’ attention. These days, they focus on research about AI in society, AI and humanity, and social media in society. They drafted an ‘AI Roadmap’ report around seven principles, and use activism, blueprinting, and lobbying to get these principles implemented. [14] [15]

Digital Humanism is a field of research and a conference series run by the Technical University of Vienna. In their words, digital humanism is “a mindset, a philosophy, a political driving force, a scientific approach, and most of all – a promise to society. That we are determined to build, regulate, and develop technology for people, for a better future”. They host a webinar series, a summer school, and a conference dealing with the subject. Digital humanism is not limited to thinking about AI, but it takes up a large portion of the issues. [16]

I go to these forums for questions rather than answers, although the latter are also valuable. But my call to action is to begin to ask ourselves questions. This goes especially to technology developers and providers. I dream of a “humanism audit” where you can find out where your development stands on serving human users, without causing harm to users, other humans, and the environment.

A last word about ethics. Traditionally, and in the business context, ethics is always related to humans; it defines norms to prevent causing harm to humans. Without a humanist approach, there can be no ethics about technology. However, I want to include the environment in the equation. Most approaches to environmentalist ethics are fundamentally anti-humanist or post-humanist; in my view, the two must be reconciled. [18] [19]

There are plenty of reasons why fully ethical technologies do not exist in our world. What we can do is take baby steps towards this, and adopting the humanist approach, and understanding what it means is a good start. [19]

Author's note

A few days after this article was sent to the editors, the humanist approach to AI and technology received more support from Pope Leo XIV, who released the encyclical letter called “Magnifica humanitas”, or “Magnificent Humanity”. In this letter, he states that “[…] the primary choice is not between a ‘yes’ or ‘no’ to technology, but rather […] between a power that claims to dominate the heavens and a people who work together in the presence of God […].” The encyclical does not necessarily make the humanist argument more valid, but it makes it more actionable. For example, conscientious refusal to use AI at the workplace could now be a legal possibility. And, of course, it has attracted a lot of attention and reactions. What to make of the presence of Anthropic co-founder Christopher Olah remains a question. [20]

*You might want to compare the humanist approach to Asimov’s Three Laws of Robotics, but those laws assume robots are sentient pseudo-persons, and our AI systems do not have this kind of sentience or personhood (yet?). In my view, present-day AI is closer to other forms of technology than to an autonomous sentient entity.