Primer: AI Nationalization
Artificial Intelligence has been experiencing a boom in the past five or so years, permeating all areas of business, government, and education. Machine learning and training language models were improving at a relatively steady pace until the end of 2022, when OpenAI released ChatGPT. The company combined Large Language Models and Generative AI into a chatbot that could respond in conversational English and complete a myriad of mundane tasks, making it extremely attractive for individuals and large organizations alike. Less than four years later, there are now 30,000 AI tech companies in the US. Currently, the largest AI companies by market value are NVIDIA (RTX), Apple (Siri), Google (Gemini), Microsoft (Copilot), and Meta (Meta AI). However, this rapid development of the AI Industry has outpaced any attempts at regulation or governance at a federal level, and many experts are sounding the alarm about the lack of AI safety.
One possible solution proposed is some form of nationalizing the AI Industry. A Politico Opinion piece by Charles Jennings details a system not unlike the Atomic Energy Commission, in which the government would “own” the AI industry but not operate it, as that task would be done by the experts, like it is now. This move would still require the relinquishing of considerable power by Big Tech companies and the consolidation of resources to create a federally run AI agency, which would entail overstepping the boundaries of the free market system and therefore be very difficult. But assuming it could be done, it could allow for greater oversight, transparency, and multidisciplinary development for Artificial Intelligence.
The reality is that AI has progressed to an extent that even its own engineers struggle to truly understand how it does what it does. This unpredictability not only makes it vulnerable to abuse by bad actors but could cause issues to major agricultural, transport, and national security systems that are presently integrating AI. A federally-run, nonpartisan agency could work to establish global leadership on AI development, better follow privacy and copyright laws, and create a centralized plan on how to make AI helpful instead of harmful. For example, ameliorating issues such as racially biased recognition systems and excessive resource use.
Those who oppose nationalizing the AI Industry argue this attempted solution would only take the somewhat divided power from Big Tech and instead place the AI monolith into the hands of whoever is in power, leading to a whole new host of problems. Entrusting the government with AI supervision and safety enforcement leaves us vulnerable to abuse of power from within, as opposed to from corporations or external threats. The same concerns about surveillance, data privacy, and biased decision-making would still remain, as administrations could tailor AI to advance their own agendas.
Considering our current situation, for example, federal AI systems would be better able to monitor communities and potentially target immigrants for deportation. However, this system could be easily manipulated into reporting any person of color at an alarming rate, increasing the severity of this particular problem.
Please come join us Monday night at 7pm in Scott Hall 201
"ChatGPT App Icon" by alpha_photo is licensed under CC BY-NC 2.0.