The launch of the Deepseek chatbot provided wake-up calls to US policymakers. “Breakthrough” appears to be exaggerated now (total spending is now underway Estimated More than $1 billion, not $6 million Originally claimedand Deepseek may have had more access Even skeptics, even more advanced tips than previously believed. Recognize That deepseek has achieved the cost-effectiveness that incumbents must now match. Therefore, Openai is quick release The O3-MINI model will be in late January 2025.
The launch of DeepSeek should not only accelerate development, but also warn policymakers who are trying to maintain US leadership in the global artificial intelligence ecosystem. You can create three policy lessons from this event.
First, it is urgent to revisit the rejection approach adopted by many Silicon Valley and the policy-making community regarding the importance of intellectual property rights. Openai's concerns The fact that despite Openai liberal leveraging other people's content, the output of that model may be used by Deepseek, indicates the myopia of this approach.
As shown in New Bookthe current weak intellectual property environment emerged from advocacy and litigation efforts by major platforms and like-minded commentators, reducing patents and copyright protections. This strategy promotes an aggregator business model that commoditizes content and technology inputs developed by others, and is then monetized within an ecosystem of complementary services.
The platform naturally prefers to minimize the price of these inputs. This can be achieved by negotiating or weakening intellectual property rights. Its legal strategies are very successful, and intellectual property rights often have little deterrent effect.
This “Light of Intellectual Property” regime has created a distorted ecosystem that has likely reduced platform and user costs, but tends to support the aggregator business model more than independent founders. For other startups with all other competitive parameters except innovation, an intellectual property portfolio can level the arena. Copyright or technical equivalents are required to maintain a market that generates payments to reflect the value of a creative asset.
In biopharmaceuticals and semiconductors, unstable intellectual property rights may have triggered venture capital. I'll shift From these patent-dependent markets that are important for public health and national security.
Second, it is essential to abandon the dogmatic “big is bad” approach to antitrust enforcement. It was pursued By the Biden administration.
The digital market tends to evolve towards higher concentration levels and therefore deserves scrutiny. Market winners do not set a barrier to rivals, but to make sure they are still competing in merit. However, this does not justify extreme skepticism about the platform adopted by antitrust enforcers in the US, the EU and elsewhere.
While intellectual property rights need to be strengthened to support the founders of technology and content in the innovation market, a targeted approach to antitrust enforcement is a key feature of the platform as a hub for the digital ecosystem. You also have to recognize the role. Platform Scale Offer perks Lower prices, lower transaction costs and expand product types.
“Rules of ReasonCourts and agencies developed during decades of federal case law should weigh these competitive “commodities” against anti-competitive “buds” caused by contested business practices. It must be.
The “Big Is Bad” approach also overlooks the US platform's global competition with its China-based rivals. Often there is a profit From state subsidies and large protected domestic markets. This is not just a commercial contest to secure market share, but a geopolitical contest that sets governance standards for information technology infrastructure underlying the world economy.
These standards ultimately reflect very different views of individual freedom and political governance – as censors show It is reportedly Specific deepseek output and reflected in regulatory authorities Concerns Through the chatbot data storage policy. This high-stakes contest between authoritarian and democratic systems takes place in both virtual markets such as AI models, operating systems, and chip designs, as well as physical markets such as data centers and submarine cables. (Inside) US-based platform is active).
Without considering the global competition facing US-based platforms, regulators can pursue dull structural relief that undermines US technical leadership. There is a precedent for this concern. The 1982 AT&T split was a groundbreaking achievement, but policymakers were unable to consider how to preserve the Bell Lab, which dumped the crown jewel assets in the US information technology industry.
Third, it is pressing for some to be rethinked Export control About the advanced AI chips proposed by the previous administration.
These rules divide the world into tier 1, 2 and 3 countries, with Tier 1 countries facing no export control, Tier 2 countries facing important control, Tier 3 countries banning exports It is facing. It is difficult to understand when limiting exports to many countries in the Tier 2 category. in Previous researchI have shown that a small number of countries will lead innovation. Placement in several countries in Tier 2, including emergency technology clusters in Israel, Switzerland and Eastern Europe, has missed technology-based cultivation opportunities Union across liberal democracy.
Although export controls in key developing markets placed in Tier 2 reflect concerns about diversion Risks of driving demand Towards “enough” alternatives from China-based rivals. Additionally, the proposed rules require that US cloud providers maintain a significant proportion of “AI calculations” in the US and tier 1 countries. That's possible I'm disappointed Providers who find geopolitical and economically important Tier 2 market facilities are: Cultivate the “Innovation Union”” Through technology-based investments.
DeepSeek's launch may disappear from attention as competitors match and outperform efficiency, price, and data security. However, it offers a valuable opportunity to act across multiple policy domains to develop a global AI ecosystem dedicated to expression, thinking and freedom of corporate freedom.
Jonathan Burnett is a law professor at Tory H. Webb at the University of Southern California, and was recently published.Big Theft: Revoking Ideology, Interest, Intellectual Property. ”
