A generation ago, Silicon Valley was known for shunning the influence game in Washington. Today, some governments maintain dedicated diplomatic posts in Silicon Valley to engage the tech firms they can no longer ignore.

That shift in power is the focus of a new paper in Policy and Society. Building on earlier work that asked why and how big tech’s power is growing, Dean’s Chair Associate Professor Araz Taeihagh from the Lee Kuan Yew School of Public Policy and his co-authors argue that big tech now wields unprecedented influence over the policy process.

“Before, we had actors that do some of these things. But we hadn't observed this happening to this degree before,” said Taeihagh. Examining this shift through the multiple streams framework (MSF), a popular approach to understanding how policy change happens, the paper coins a term for actors in this new, superlative policy role: ‘super policy entrepreneurs’, or SPEs.

SPEs’ expanding super-influence

Although big tech is of course not a monolith, the authors argue that most tech giants have diverse tools and strategies at their disposal, and disproportionate influence on the policy process. Given big tech’s rising power, Taeihagh and his colleagues had expected to see more clashes between these firms and governments. Instead, over the last few years, the growing popularity of artificial intelligence (AI), particularly generative AI (GenAI), has shifted the two parties towards “symbiosis”, as Taeihagh put it, with the normalisation of tech tools for achieving policy goals.

GenAI’s wide-ranging use cases make it highly attractive for everyday government operations. But most leading models are owned or controlled by big tech, which creates a ‘self-perpetuating cycle’ where each new tech breakthrough serves to further cement big tech’s market and political dominance. In the first half of 2025, noted Taeihagh, roughly half of America’s inflation-adjusted gross domestic product growth was driven by AI-related capital spending.

What this means for the rest of us

Conversations around tech have become “technocratic”, Taeihagh commented, limiting decision-making power to a few technical experts, instead of more participatory engagement. Without broader public participation, he said, societies may be less prepared to guard against the social harms of these technologies – from scams and deepfakes to the erosion of local culture.

On the global stage, the power imbalance is especially stark in developing countries, where governments have little bargaining power against tech firms whose services and infrastructure they would be hard-pressed to do without. That dependence gives big tech substantial influence over local economies and policymaking.

Good news and bad news

The good news: States and public institutions are far from passive, and public backlash can still block big tech from getting their way. Thus, “it is very important for everyone to understand what is going on,” said Taeihagh.

The bad news is that the intensifying AI race between the United States and China is pushing both sides into a race to the bottom, making governments less willing to regulate their own tech firms for fear of falling behind. To mitigate this, Taeihagh argues for a ‘minimum viable agreement’ between major powers to govern the most dangerous uses of AI. Outside of the major powers, he advocates for greater regional cooperation, so that countries can negotiate more effectively as blocs rather than individual states.

Despite his dark warnings, Taeihagh remains hopeful about the future of AI development. “People understand that this concentration of power is a problem,” he said. “It's just that the speed of development is so fast, making sure we don't get left behind is the challenge.”