"The Ideological Struggle Behind AI: More Than Good Versus Evil"

This article is a summary of a article in Dutch written by Siri Beerends & Emile van Bergen here you can find the original tone. https://www.setup.nl/artikelen/de-existentiele-dreiging-van-ai-de-ideologie-erachter/

Ethical AI: a contradiction

A tale of AI protagonists and AI antagonists. The prevailing notion is, "If the wealthy elite behaves ethically and doesn't rush too quickly, the development of AI will turn out fine." The belief is, "We just need to establish appropriate rules to guide the progress of AI." While the appeal of this idea and the hope for 'ethical AI' are understandable, it constitutes an inherent contradiction.

AI is not just a matter of developing 'good' or 'bad,' or responsible versus irresponsible' use. Ethically, AI is inherently problematic and ideologically charged.

1. Not everything that is important to humans can be captured in data. However, by attempting to do so, we distort reality—either by denying the existence of the lived experience or by adapting human perception to fit the AI model.
2. AI instrumentalizes intelligence for economic purposes. Our human intelligence, which we use to produce texts, images, and other data, is reduced to raw material to train AI. The resulting machine 'intelligence' serves the interests of the provider, not the user.

It seems like AI is rapidly advancing, but in reality, AI has nothing to do with the increase in intelligence. Instead, it is driven by the increase in data, computational power, and the diligence to feed AI applications with texts, images, and more.

Data and parameters are not synonymous with insight, understanding, or reasoning. From the data and parameters, the most probable answer is generated through probability calculations.

AI researchers often possess a simplified understanding of language, human processes, and evolutionary complexity. Existential threat scenarios do not take into account the chaotic nature of the natural world. The structures of humans and computer programs are fundamentally different, resulting in no shared experiential world and meaning.

reflection: As humans, we often seek to model and simplify everything to grasp it more easily, but in doing so, we risk losing nuances. How can we truly derive meaning if we ignore these subtleties?

A computer with an AI program is not an independent life form and cannot exist without humans. The actual risk lies with us, as we have the choice to stop, redirect, or proceed. The threat of AI does not lie in sudden disasters but in gradual processes of decline. AI that poses a threat to humans is often viewed as a future risk, but current AI already threatens us existentially—not physically, but by reducing us to economic functions.

To make complex situations understandable, we try as humans, to put processes, knowledge, systems into visuals, models, frameworks, matrices etc… At best, these models are simplifications of reality and a stepping stone to a deeper understanding of the whole. But sometimes they are just too simple and too one-sided and they go awry. Yet they remain common good. one of these widely used models is this of an organization system.

?

If the robot doesn't resemble a human, then we turn a human into a robot.

Humans are increasingly becoming like machines. By reducing language comprehension, consciousness, intelligence, perception, intuition, empathy, and personality to statistical phenomena and data processing, AI companies can assert that they can replicate these aspects with their own AI systems. The overestimation of AGI and x-risk impacts the human self-image and the way we understand and act upon intelligence in our daily lives.

?A study (the Impact of AI assistance on student agency conducted by Ali Darvishi, Hassan Khosravi, Shazia Sadiq, Dragan Gaˇsevi′c, and George Siemens.) focusing on providing AI support for writing feedback, suggests that students tended to rely on AI assistance rather than actively learning from it.

The Complexity of Intelligence and Humanity: A Philosophical Exploration

The concept of intelligence and the nature of being human have been subjects of philosophical debate for centuries. Entrusting these immense questions to a group of computer enthusiasts may seem absurd at first glance.

Intelligence is an abstract, versatile concept for which there is fortunately no universal, unambiguous definition. Terms like "at human level" or "beyond human level" suggest a measurable standard of humanity. In reality, intelligence is culturally determined and extends beyond what we can ascertain with Western measurement tools and simplistic IQ and EQ tests.

Let us not forget that in history, entire communities were exterminated both physically and mentally in the name of intelligence and the question of whether they were human or not."

The AI industry often models artificial intelligence according to human standards, indirectly determining the definition of a human. However, we must keep this definition open. The ambiguity of intelligence is not a shortcoming but a necessary condition for being human. The fact that we cannot fully understand and describe our human capabilities offers us the opportunity to be part of an open future.

Unfortunately, this open future is currently under threat. As intelligence is narrowly and mechanistically defined, AI companies gain the power to shape our understanding of intelligence according to their own commercial interests. The tragic consequence is that we reduce the meaning of intelligence and being human to what AI developers can or cannot replicate.

The threat of extinction does not come so much from a runaway AGI but rather from our own tendency to act increasingly mechanically. Let us not fear so much the machine that suddenly grows outside of us but especially the machine that gradually grows within ourselves.

This article is a summary of a article in Dutch written by Siri Beerends & Emile van Bergen here you can find the original tone. https://www.setup.nl/artikelen/de-existentiele-dreiging-van-ai-de-ideologie-erachter/

references:

the Impact of AI assistance on student agency conducted by Ali Darvishi, Hassan Khosravi, Shazia Sadiq, Dragan Gaˇsevi′c, and George Siemens.

BOBIP part 1: How do we, as humans, think organizations work and what happens in reality? - Evelien Verschroeven