In a fast-changing world, where countries are racing to become the leading force in artificial intelligence (AI), former Google CEO, Eric Schmidt has issued a stern warning to Western nations to consider open-source AI. He made the call during the AI Action Summit in Paris, as various stakeholders discuss the benefits of open-source vs. proprietary AI models. This discussion gains more significance following the emergence of Chinese company DeepSeek and its recent development of an AI model.
The emergence of Deepseek
A New Player on the Global Stage DeepSeek made a shock entry in the market with the R1 model, which has left many industries and the U.S government stunned. The R1 model of DeepSeek is an open-source model with a focus on efficiency and strong reasoning. This is in sharp contrast with most of the AI models in the USA, which are currently proprietary. Some of the major AI models such as Google’s Gemini, Anthropic’s Claude, and OpenAI’s GPT-4 are proprietary and are only available to those who can pay the exorbitant costs of the license fees, which is used to subsidize the astronomical training and infrastructure costs.Deepseek is a free to use service with its reported ability to achieve comparable performance to leading AI models while utilizing fewer advanced chips and significantly lower training costs. In light of this, Schmidt’s warning is quite relevant: if the West fails to switch to open-source AI, the market will be captured by countries such as China that are fast developing open-source AI.
The Open Vs. Closed Debate
The contest between open source and closed source, which is the basic concept of this debate, needs to be explained. Open source models, by their very nature, are such that researchers, developers, and academic entities can easily obtain, modify, and extend latest advancements in technology. This openness has the potential of fostering a community that can speed up research and creation. On the other hand, closed source models, although they can be used to create high performing and secure products, are deployed in a manner that makes it easier for a few highly funded organizations and companies to control the flow of information and progress. Schmidt has said that the West has spent a lot of money on closed systems and this may have some implications that the West cannot control.
“If we don’t do something about that, China will eventually become the open source leader and the rest of the world will become closed source”, he told the Financial Times. This scenario may limit the development of science and technology in Western countries and businesses, which may not be able to afford the expensive closed systems.
The Economic and Scientific Implications
Open and closed AI models are not just technological choices; they have important economic and scientific consequences. Schmidt pointed out that although the U.S. has been at the forefront of the initial phase of AI development with achievements such as GPT-4, Europe is still in a position to seize the opportunity by focusing on what he called the “application layer.” This approach is based on the use of AI to enhance productivity across multiple industries. “Europe needs to shape up,” Schmidt said, emphasizing that there is an opportunity for Europe to gain an economic benefit from a strong open source AI system. The investment dynamics further complicate this landscape. The largest tech companies in the USA including Google, Open AI and Amazon have invested billions of dollars in the production of proprietary AI models and this has been seen by many as a way of generating returns on investment. But, there is a voice within this paradigm which is saying that the balance has to be struck. OpenAI’s CEO, Sam Altman, has also lately commented on the company’s strategy, stating that perhaps he was on the wrong side of the world on this one by paying too much attention to proprietary models. This is because Altman’s comments bring out the advantages that come with open source models as much as they present new problems.
The Call for Collaboration and Safety
However, there is another critical aspect of the argument, which is the safety implication. Besides the competitive aspects of the debate, Schmidt also highlighted the safety aspects of the research in AI. He unveiled a $10 million fund for AI Safety Science program under Schmidt Sciences, a not-for-profit organization he established with his wife Wendy. This initiative will support 27 projects that are meant to tackle some of the biggest safety problems in AI, with participants like Turing award winner Yoshua Bengio and Carnegie Mellon professor Zico Kolter. In a very important call for international cooperation, Schmidt proposed that the West should partner with China on AI safety.
“How could it possibly be bad for us to give them information that they could use to make their models more safe?” he asked in comparison to military rocket testing which is collaborative. This viewpoint stresses that, although there are stiff competition among the various parties involved, there are still some issues that are of concern to all, such as the safety of the new and rapidly developing AI platforms.
Looking Ahead: A Hybrid Future?
As the open source versus closed source models debate gains more momentum, more and more people are calling for an intermediate model that incorporates both principles. For his part, Schmidt has said that AI dominance may be achieved through a mix of the two approaches. This integrated model would permit the development of strong and secure proprietary systems, as well as the open source cooperative and versatile innovation. As the world enters the new era of the AI race, the decisions that Western countries will make in the next few years will prove to be critical.
