Governments should focus on AI adoption, not just regulation

Millions of voters are going to the polls this year. New administrations in charge are going to be faced with questions about Artificial Intelligence (AI). Most are focusing – and will focus – on legislation. But how many governments will use AI to better carry out their business?

The AI frenzy has sent governments “racing to control their technological destinies,". At the 2024 World Governments Summit, the CEO of NVIDIA blessed the race by admonishing spectators that they should build their “sovereign intelligence.” Last year ended with the European Union agreeing to the AI Act, and this year has started with bets being placed on where the next AI regulations and policies will emerge.

Should we regulate AI? Do we need more public investment in AI? These are essential dilemmas, but they miss an elephant in the room. They only see governments as producers of policies or shapers of markets. Yet we should see governments as adopters of technology. According to the OECD, the most productive firms are those who are better at AI adoption. Why not government?

Governments should see themselves as their own customers in AI. Civil servants are asking for it. Research conducted by Apolitical (a company I work with) shows that most of the government workforce is optimistic about Generative AI. In fact, the number of users in government has been steadily growing, but behind the scenes.

Today, institutions rely on enthusiasts who experiment with AI at their own risk because they lack clear guidelines for safe and secure trials. "Shadow AI" is, at best, an incomplete plan, a missed opportunity for governments to raise their operational excellence. At worse, it can expose administrations to vulnerabilities and misuses.

To be fair, the use of AI in government is not new. Countries like Estonia have a tradition of using AI for tens of government applications. What is new with Generative AI is how accessible and easy to use it is. You do not need to be a prompt engineer to start benefiting from Large Language Models. And “the AI you are using now is the worst AI you will ever use,” as Professor Ethan Mollick of the Wharton School often writes.

As Generative AI inevitably spreads from IT departments to everyone in civil service, what should governments do?

To begin with, clarity is kindness. Users are likely to face analysis paralysis if institutional guidelines are unclear or burdensome. They need use cases, validation, and support networks to trial the tech, safely and securely. Clarity yields confidence and, in AI, as in management or investment, a predictable environment is a productive environment.

Governments need to build enabling environments more than prescriptive guidance. As tech expert Azeem Azhar puts it: “A major challenge for any leader is balancing the need to grab the opportunity of generative AI, but also contending with the fundamental unknowability of the technology.” The UK Government recently published a Generative AI Framework that moves in that direction.

Clarity is necessary but insufficient without investment in competencies. GovTech Singapore created a Data&AI literacy primer, which has been completed by 60% of its workforce. They tailored learning paths to the needs of public servants, worked flexibly with departments, and built communities of learners. Other governments should follow.

I am not advocating for Generative AI to start reigning over government departments. When too much of a new technology is released too quickly into critical functions, the consequences can be dire. But countries that came too late to the innovation party during the digital age went from being an economic powerhouse to a stasis. In the AI age, the downfall could be much quicker.

Both governments and AI are here to stay. From my time in tech, I learnt that “you can build a business strategy around the things that are stable in time." If governments want to truly take control of their AI destinies, they should use their tech. Else the opposite may happen.

MPP alum Leonardo Quattrucci advises governments and businesses in putting technology to public service. He has over a decade of experience in building innovative programmes across tech and the public sector, from the European Commission’s in-house think tank to Amazon Web Services’ Center for Quantum Networking. His work has been recognised by Forbes, BMW Foundation, World Economic Forum, and the Aspen Institute. Leonardo writes 6 Minutes Well Spent: a newsletter about overlooked perspectives in tech. Find him on LinkedIn or X (formerly Twitter).