Last week, the Wall Street Journal claimed that OpenAI CEO Sam Altman was planning to raise as much as $7 trillion to transform the world’s semiconductor sector into an AI powerhouse. Another evidence of the market concentration in generative AI is the fact that a single startup could present a funding goal higher than Japan’s GDP and still not be laughed at out of the room.
There are too many reasons why artificial intelligence (AI) shouldn’t be monopolized:
- Its potential benefits to medicine
- The risks of election meddling
- The prospects for useful climate research
- The challenge of solving basic physics problems
However, this is precisely the trend the industry is showing, as only a handful of organizations have the means and expertise to create the most cutting-edge AI. The training of massive language models for various AI applications necessitates data and processing capacity that are particularly taxing on resources. If they don’t break free of their reliance on Big Tech, researchers and SMEs will once again fall behind in the innovation race.
Public funds are being poured into efforts to equalize computational power on both parties of the Atlantic. It was last month that the US government launched the National AI Research Resource to guarantee researchers had access to resources on the same level as Silicon Valley behemoths. The US National Science Foundation is spearheading this experimental study. The education and research community will be able to use government-funded data and computation to construct and comprehend AI, thanks to its collaboration with ten other federal agencies and twenty-five civil society groups.
Before the current surge of generative AI, the European Union established a decentralized network of supercomputers in 2018 with a comparable goal. It would appear that the EuroHPC was underutilized and has been living in oblivion. Ursula von der Leyen, president of the European Commission, stated at the end of last year that we must utilize this power. The European Union currently thinks that democratic access to supercomputers can also aid in the establishment of “AI factories,” in which tiny companies combine their resources to create innovative models.
As a public utility, internet connectivity has long been discussed due to its significance in many aspects of society, including education, employment, and information acquisition. But regulations aimed at achieving that goal were never accepted. However, after computing was declassified as a public utility, the United States and the European Union showed a genuine desire to invest in public digital infrastructure.
New policies may just be industrial policy revamped, but they are finally helping to control the digital economy and limit the huge impact of IT companies. Governments have taken the right call by increasing access to basic computing tools, but this is just the beginning; further action is required in the form of legislation and regulations. It is the responsibility of antitrust agencies to prevent the biggest AI businesses from becoming uncontrollably large. It is the responsibility of security authorities to block criminals from gaining access to vital computer resources.
AI applications exhibit bias and discrimination in a variety of ways, which is a major concern for anti-discrimination advocates. In the same way, public investments in AI are helping regulations that aim to stop market monopolies from developing into knowledge monopolies. Even though the European Union’s Digital Services Act, which details the obligations of platform providers, wisely encodes academics’ access to data, the AI Act does not contain any such language. For instance, businesses must disclose their data inputs and energy consumption, but trade privacy will be upheld, enabling considerable suppression of crucial facts.
Moving ahead, public digital infrastructure investments need to ramp up, and state funds should be shifted away from Big Tech, regardless of whether the projects serve a public purpose. Though a considerable figure, the US government’s $3.3 billion investment in AI in 2022 looks small in comparison to the industry’s tens of billions of dollars or Altman’s trillions.
A better public knowledge of AI makes the prevention of AI monopolies all the more important, and it is an integral aspect of a flourishing innovation atmosphere. There is some overlap between those objectives here. Many important advancements in technology have their origins in academic research. We must ensure that the environment is not strangled.
