Regulators Should Not Be Intimidated by the "Dominant AI Forces"
In a recent op-ed published in Wired, Bhaskar Chakravorti, dean of global business at Tufts University's Fletcher School of Law and Diplomacy, argues for U.S. antitrust regulators to take preemptive action against the growing concentration in AI. However, the debate surrounding this issue is far from settled, with concerns about the potential risks to American competitiveness and innovation in AI.
According to Chakravorti, the vast majority of innovations in AI are beneficial for society and pose little risks. Yet, many of the targets of the antitrust movement are large tech companies that tend to use AI, leading some to question the necessity of such actions.
The top 500 AI patent applicants globally include 167 universities and public research organizations, as well as many businesses outside the tech industry, such as Capital One, Bank of America, and Accenture. In 2019, U.S. companies accounted for 64% of the total global private investment in AI.
Critics argue that ongoing and potential antitrust suits lack clear evidence of consumer harm and may inadvertently disrupt the financial and innovation engines that sustain U.S. leadership in AI technology. Defining a distinct market for AI products is legally and economically challenging, with ongoing debates about data access asymmetries and platform gatekeeping but without settled standards.
One of the main concerns is the risk to U.S. AI dominance. Big tech firms are investing hundreds of billions in AI research and development, funded by revenues from existing business models. Suits that weaken these revenue streams could hobble the financial basis for AI innovation, putting the U.S. at a disadvantage globally.
Another concern is the potential chilling effect on innovation. Heavy-handed regulation may disincentivize investment and talent acquisition in AI, slowing progress at a critical moment of technological competition. There are nuanced concerns about how AI systems use nonpublic data and pricing algorithms, with courts divided on cases alleging anticompetitive algorithmic practices, showing the difficulty of drawing bright lines for regulation.
Chakravorti proposes alternative strategies to promote responsible AI development, such as closer scrutiny of acquisitions of AI startups by major tech companies by antitrust regulators, using tax policy to incentivize investments in "value-enhancing" AI instead of "value-destroying" AI, and establishing a 'creative commons' for AI R&D, including mandating open IP for AI patents.
However, the author does not provide specific details on what closer scrutiny of AI startup acquisitions by major tech companies would entail. The top 30 U.S. companies only held 29% of all AI patents granted between 1976 and 2018, suggesting a diverse landscape of AI innovation.
In light of these arguments, the debate over preemptive antitrust enforcement in AI is complex and multifaceted, with concerns about both protecting American competitiveness and fostering innovation on one hand, and ensuring fair competition and consumer protection on the other. As the AI market continues to evolve, it is crucial for regulators to carefully consider their actions to strike a balance between these competing interests.
Two of the most popular machine learning libraries, Tensorflow (created by Google) and PyTorch (created by Facebook), are available as open source, further demonstrating the collaborative nature of AI development.
- The debate over preemptive antitrust enforcement in AI is complex and multifaceted, as it involves protecting American competitiveness and fostering innovation, while ensuring fair competition and consumer protection.
- In light of this, it is crucial for regulators to carefully consider their actions to strike a balance between these competing interests.
- Two of the most popular machine learning libraries, Tensorflow (created by Google) and PyTorch (created by Facebook), are available as open source, further demonstrating the collaborative nature of AI development.
- Chakravorti's proposals for promoting responsible AI development include closer scrutiny of acquisitions of AI startups by major tech companies, using tax policy to incentivize investments in 'value-enhancing' AI, and establishing a 'creative commons' for AI R&D, including mandating open IP for AI patents.
