Artificial Intelligence: The University’s Losing Match with the Private Sector

laboratory life. “Such growth, in such a short time, has never been heard in any field”, observes Neil Thompson. This specialist in artificial intelligence (AI) and economics at the Massachusetts Institute of Technology (MIT, Boston) is not talking about the crazy growth of users of the ChatGPT talking machine developed by OpenAI. Nor has the pervasive presence of the concept of artificial intelligence in the media since the publication, in November 2022, of this conversational agent and its derivatives or applications. It leads to a deeper and more disturbing fact: the apparent imbalance between the academic world and the industrial world in the field of artificial intelligence. The other folded the game in just ten years.

In American Journal of Science ScienceOn March 2, with Noor Ahmed, also at MIT, and Muntasir Vahed (Virginia Tech University, Blacksburg), Neil Thompson quantified this win. In 2020, the vast majority of AI PhD students in the United States, 70%, were hired by the private sector, compared to only 20% in 2004. The researchers also found that poaching of professors by large companies increased by 10% in 2018, while university hiring remained stable for fifteen years.

On the computing resources front, the same imbalance. Industry “models”, that is, programs that have been prepared on a huge scale, such as ChatGPT, Bard, Dall-E, etc. than in the academic world. However, in many cases, larger size is synonymous with higher quality.

As a result, in 2020, 40% of presentations at conferences came from private laboratories, which is more than double what they represented in 2012. On this topic, in April, their colleagues in the Stanford annual report made similar observations: the industry had thirty-two models. “important” against three in the academic world. As for publications, the report notes that even on the ethical aspect of artificial intelligence, companies “Send More Than Ever”. For example, in 2022, three times more than in 2021, and their number is equal to a third of the production of the academic world.

“Even at the Massachusetts Institute of Technology we can’t fight anymore”

in short, “According to the three key success parameters of AI models, data, computing power and human resources, it is unbalanced”Summarizes Noor Ahmed, who recalls that companies will spend $340 billion on artificial intelligence in 2021, when US funding agencies support research for $1.5 billion that year. “Even at the Massachusetts Institute of Technology, you hear colleagues say they can’t fight anymore”Testified by Neil Thompson.