How AI writes our laws

Second, we need to strengthen disclosure requirements on lobbyists, whether fully human or AI-assisted. State laws governing lobbying disclosures are a hodgepodge. For example, North Dakota only requires lobbying reports to be filed annually, so the policy may already be determined by the time disclosure is made. An official lobbying scorecard created by Open Secrets, a group that studies the influence of money in American politics, tracks nine states that don’t even require lobbyists to report compensation.

Ideally, it would be great for the public to see all communications between lobbyists and legislators, whether or not it takes the form of reform. In the absence of that, let’s give the public what to judge Lobbyists for—and why. Lobbying is traditionally an activity behind closed doors. Currently, many states reinforce that: they exempt testimony given publicly to the legislature from being reported as lobbying.

In those jurisdictions, if you’ve made your position public, you’re no longer lobbying. Let’s do the opposite: ask lobbyists to state their positions on issues. Some states require a position statement (‘yes’ or ‘no’) in advance from registered lobbyists. And in most (but not all) states, you can file a public records request for meetings with the state legislature and hope to get something useful back. But we can expect more—lobbyists may be expected to proactively publish within days a brief summary of what they’ve asked policymakers for during meetings and why they believe it’s in the public interest.

We cannot trust corporations to be forthcoming and completely honest about the reasons behind their lobbying positions. But keeping them on record would at least be a starting point for accountability.

Finally, consider the potential role AI assistive technologies will have on the labor market for corporations themselves and lobbyists. Many observers are rightly concerned about the potential for AI to replace or reduce manual labor. If the automation potential of AI ends up improving the work of political strategy and message development, it will put some professionals on K Street out of a job.

But don’t expect this to hamper the work of the astronomically paid lobbyists: former members of Congress and other insiders who have passed through the back door. There is still no shortage of reform proposals to limit the ability of government officials to sell lobbyists to their colleagues in government, and they must be adopted and — just as importantly — preserved and implemented in successive Congresses and administrations.

None of these solutions are specific to the threats posed by AI, or focused primarily on micro-law — and that’s the point. Good governance should and can be resilient to risks from different techniques and actors.

But the risks caused by AI are especially pressing now, how fast the field is growing. We expect the scale, tactics, and effectiveness of human lobbying to evolve over the years and decades. Advances in AI, on the other hand, seem to be making amazing breakthroughs at a very rapid pace – and it’s still accelerating.

As the legislative process is amended, written, and expanded at the federal, state, and local levels, it is a constant struggle between parties seeking to control the rules of our society. Lobbying is an important tool for balancing different interests through our system. If done well, perhaps lobbying can help policymakers make fair decisions on behalf of all of us.

Nathan E. Sanders He is a data scientist and associate with the Berkman Klein Center at Harvard University. Bruce Schneier He is a security technology expert and a fellow and lecturer at the Harvard Kennedy School.