On May 19, 2025, Jensen Huang, the founder and CEO of Nvidia, announced that his company would build its new headquarters right next to my high school^[1]. Just like that, he forever changed the landscape of my town, now he will forever change the landscape of our human condition.

Jensen Huang and many of his fellow AI CEOs have been giving a promising image for the future of humankind with their progress on Artificial Intelligence (AI) and the coming of Artificial General Intelligence (AGI). Sam Altman wrote, “the gains to quality of life from AI driving faster scientific progress and increased productivity will be enormous; the future can be vastly better than the present……[and] the 2030s may not be wildly different. People will still love their families, express their creativity, play games, and swim in lakes.^[2]”

However, oddly enough, when giving a commencement speech^[3] at National Taiwan University, to the winner of the meritocracy system, Jensen Huang did not reassure the graduates of their future with AI. Instead, he explicitly warned the graduates “to run but not to walk.” Stating that if they don’t work hard enough, they would be replaced by AI.

It made me wonder why those tech CEOs are presenting a wonderful utopian future on one hand, but in the meantime, telling everyone to work harder to not be replaced. The question hence raised — What future is AI actually bringing? And how will the impact of AI affect the human condition?

I think AI is going to bring us to a dystopian world. While many analyses examine AI from single perspectives, and campus discussions often reflect vague optimism, this essay offers a critical overview drawing on conversations with AI industry insiders. As someone from Taiwan,  constantly surrounded by AI engineers, I aim to provide a grounded entry point for understanding AI’s multifaceted impacts on our future.

First, I will examine Ai’s impact on the individual and its likely influence on our labor conditions. Then, I will explore how the production process of AI knowledge is influenced by power dynamics and is designed to achieve specific future goals. Finally, I will discuss potential solutions and actions we can take to address the challenges posed by AI.

AI in our life

Recent MIT media lab research^[4] found that using Large Language Models (LLMs) like ChatGPT for essay writing results in an accumulation of cognitive debt. The study demonstrated that brain connectivity systematically scaled down with increased external support, showing the weakest overall neural coupling in the LLM-assisted group, indicating that cognitive effort was offloaded to the tool. This LLM usage resulted in statistically homogenous essays, a significantly impaired ability to quote correctly from their own writing, and a low perception of ownership. These findings collectively indicated a likely decrease in learning skills among those consistently relying on LLMs.

          Beyond individual cognitive impacts, AI is reshaping labor markets. The New York times^[5] wrote that the job market for economists is over and that recent graduates in the US cannot find a job. See below figure by Oliver Kim^[6]. This is the situation in the US, but the UK is not much better. Research at King’s College London^[7] found a 4.5%reduction in total employment at firms with high exposure to AI. They found the AI effect is predominantly concentrated in junior positions, with a decline of 5.8%.

Despite how mainstream economists’ rationale is less and less reliable in this dystopian world^[8], they still conclude that “advances in technology (specially, in artificial intelligence and robotisation) are poised to make such massive increases in labour productivity that vast swathes of the current workforce, far from being asked to work for longer, will find themselves redundant because there is no longer enough work for human beings to do^[9]”

How have we arrived at AI?

          These labor effects stem from deeper power dynamics in AI development. When we look into the power dynamics of the production of AI knowledge, we will find that it is highly centralised with a very specific purpose. The Special Competitive Studies Program (SCPS) wrote explicitly in their memo to the president^[10] that “the future of global power and prosperity hinges on the United States’ ability to” maintain the leadership at AI competition and to be first in developing the AGI to “secure dominance for the next century.”

          The RAND Company’s research^[11] found that the market for AI foundation models currently exhibits characteristics of a natural monopoly, where a single firm can serve demand more efficiently than multiple competitors. This stems from high economies of scale, substantial sunk costs in computing infrastructure and R&D, as well as significant economies of scope as firms reuse foundation models across product lines. The required investments in data acquisition, labor, and initial computing create formidable barriers to entry that reinforce market concentration.

Cecilia Rikap argues^[12] that this concentration is problematic, framing Big Tech companies as intellectual monopolies controlling Corporate Innovation Systems (CIS). These firms systematically capture and transform knowledge into rent-bearing assets, often co-produced with universities and research institutions. This monopolization has negative societal outcomes: technology deployment that increases inequality, weakens workers, and fails to create shared prosperity. Big Tech’s centralized control allows them to steer scientific exploration and technological trajectories through their hub positions in the global innovation network. For instance, Big Tech executives dominate AI conference organizing committees, effectively deciding which research gets presented and awarded, thereby shaping the field’s frontier while curtailing knowledge diffusion and stifling complementary innovations that could distribute technological benefits more widely.

What can we do?

          Understanding these structural forces reveals why AI is unlikely to be the savior, as claimed by tech CEOs. Instead, it serves as a tool for those in power to enhance and secure their dominion. Still, I believe that there must be something we can do. Rikap’s campaign for her program on building South technology infrastructure^[13] to help people seize back the technological means of production. While Ambassador Tang^[14] at Oxford reminds us that the AGI has already existed  around us.

          The Artificial Intelligence revolution started 2017 with a paper named Attention is all you need^[15]. However, I believe that Artificial Intelligence is not about you — the researcher, the tech CEOs, but about everyone who lives in this world that is so heavily shaped by Artificial Intelligence with or without our consent. In conclusion, Attention is all we need. If we all start thinking about our future with Artificial Intelligence, to stop being blindly optimistic about AI and start acting for a better future, we could find a way out for the upcoming dystopia.

Bibliography 

[1] Monica Chen Asia Taipei; Charlene Chen, DIGITIMES, “Nvidia Bets Big on Taiwan as New HQ Takes Shape, Pending Final Land Talks,” DIGITIMES, May 19, 2025, https://www.digitimes.com/news/a20250519PD231/nvidia-jensen-huang-taipei-taiwan-2025.html.

[2] Sam Altman, “The Gentle Singularity,” Sam Altman, 2025, https://blog.samaltman.com/the-gentle-singularity.

[3] Jensen Huang, “Jensen Huang’s 2023 National Taiwan University (NTU) Commencement Speech,” YouTube, 2025, https://youtu.be/oi89u6q0_AY?si=VM8PAqQoGinlcQVl.

[4] Nataliya Kosmyna et al., “Your Brain on ChatGPT: Accumulation of Cognitive Debt When Using an AI Assistant for Essay Writing Task,” version 1, preprint, arXiv, 2025, https://doi.org/10.48550/ARXIV.2506.08872.

[5] Noam Scheiber, “The Bull Market for Economists Is Over. It’s an Ominous Sign for the Economy.,” Business, The New York Times, July 28, 2025, https://www.nytimes.com/2025/07/28/business/economics-jobs-hiring.html.

[6] Oliver Kim, “Twilight of the Econs?” January 21, 2025, https://www.global-developments.org/p/twilight-of-the-econs.

[7] Bouke Klein Teeselink, “Generative AI and Labor Market Outcomes: Evidence from the United Kingdom,” preprint, SSRN, 2025, https://doi.org/10.2139/ssrn.5516798.

[8] Ismail Abdi, ‘Rational Economics’ in a Dystopian World? • London Geopolitical Review, September 21, 2025, https://londongeopoliticalreview.co.uk/rational-economics-in-a-dystopian-world/.

[9] Tim Jackson, “The Post-Growth Challenge: Secular Stagnation, Inequality and the Limits to Growth,” Ecological Economics 156 (February 2019): 236–46, https://doi.org/10.1016/j.ecolecon.2018.10.010.

[10] Special Competitive Studies Project, Memos to the President: Artificial General Intelligence (AGI), Scsp.Ai (2025), https://www.scsp.ai/wp-content/uploads/2025/01/AGI-Memo.pdf.

[11] Jon Schmid et al., Evaluating Natural Monopoly Conditions in the AI Foundation Model Market (2024), https://www.rand.org/pubs/research_reports/RRA3415-1.html.

[12] Cecilia Rikap, “The US National Security State and Big Tech: Frenemy Relations and Innovation Planning in Turbulent Times,” Edward Elgar Publishing eBooks, Edward Elgar Publishing, 2025, 74–90, https://doi.org/10.4337/9781035357178.00009.

[13] Cecilia Rikap et al., “Reclaiming Digital Sovereignty: A Roadmap to Build a Digital Stack for People and the Planet,” in UCL Institute for Innovation and Public Purpose (IIPP): London, UK., Report (UCL Institute for Innovation and Public Purpose (IIPP), 2024), https://www.ucl.ac.uk/bartlett/public-purpose/publications/2024/dec/reclaiming-digital-sovereignty.

[14] Audrey Tang and Caroline Green, “6-Pack of Care: A Manifesto,” 6pack.Care, 2024, https://6pack.care/manifesto/.

[15] Ashish Vaswani et al., “Attention Is All You Need,” version 7, preprint, arXiv, 2017, https://doi.org/10.48550/ARXIV.1706.03762. For a more detail explain for the paper and why it is important https://en.wikipedia.org/wiki/Attention_Is_All_You_Need