Unlocking AI’s Potential for Everyone
At a time when antitrust authorities are already growing concerned about Big Tech’s market power, the industry’s dominant players are now deploying artificial-intelligence models to reinforce their position. To ensure that these powerful new technologies serve everyone, policymakers must set some basic rules of the road.
CAMBRIDGE – Artificial intelligence is moving fast. People are using generative AI and large language models (LLMs) to build new services and perform existing tasks, and the underlying technology itself is advancing quickly. As the Nobel laureate economist Michael Spence observes, this wave of adoption could well yield significant productivity gains, after almost two decades of lackluster growth. Every day brings news like Google’s recent announcement that its AI has helped American Airlines reduce contrails by 54%, reducing each flight’s climate footprint.
But the news isn’t all good. As matters stand, AI is more likely to help Big Tech companies cement their dominance. They are the ones with the resources to develop and maintain the most powerful AI models, and they are already moving quickly to bundle LLMs with their existing services. These developments come at a time when antitrust authorities around the world are already growing increasingly concerned about tech companies’ market power.
To be sure, some commentators – including one Google engineer, in an internal memo – argue that this fear is overblown, owing to the presence of open-source LLMs that technically allow for anyone to compete in the market. But even if there is a blossoming of smaller new entrants, Big Tech’s dominance still looks secure. A recent paper comparing open-source models to the AI application programming interface (API) services that Big Tech companies are providing to third parties finds that the latter perform much better on most criteria.