Large language models (LLM) such as ChatGPT have the potential to shake up political lobbying according to research conducted at one of the top US Universities.
An investigation by John Nay of Stanford University posits that at least part of the lobbying work from corporate interests can be automated. Nay argues that this development could be good news for politics, but also warms that there are inherent dangers. Chatbot technology could even be leveraged to undermine the integrity of political institutions and the legislative process.
Lobbying for policy change
A recent paper by John Nay of Stanford University titled Large Language Models as Corporate Lobbyists suggests that chatbots such as ChatGPT have the potential to become competent political lobbyists.
To test the theory Nay prompted ChatGPT to complete a set number of diverse tasks necessary for a functioning lobbying process.
Nay first asked the system to determine whether proposed US Congressional bills are relevant to specific companies. The chatbot was further asked to provide reasons for its assessment of the legislation and to offer confidence levels in its own analysis.
For the bills ChatGPT considered relevant, the bot was then asked to draft a lobbying letter to the bill’s sponsor and to ask for changes to the proposed legislation.
The research, which was published earlier this month, found that while ChatGPT is still some way from being able to fully replace the lobbying efforts of a human being that this situation may change in the future.
The report went on to say that “as large language models continue to exhibit improved natural language understanding capabilities, performance on corporate lobbying related tasks will continue to improve.”
According to Nay the benefits of AI as lobbyists are twofold. One is that it allows humans to spend less time on mundane tasks freeing them to focus on challenges at a higher level. The second is that it makes lobbying easier and more affordable, opening the practice to a wider number of potential players.
Could ChatGPT negotiate on your behalf?
Nay argues that ChatGPT could negotiate on behalf of the little guy and there is something fairly compelling about this pitch.
On the other hand, ChatGPT and other chatbots could instead be leveraged by large corporations and other well-resourced players to further extend their advantage in the political sphere.
Rather than improving our democracies, chatbots could instead flood them with a wave of increased corporate lobbying, eroding the relative power of ordinary people.
Besides lobbying directly on bills, chatbots could be further utilized to lobby individual politicians, write email campaigns, or game social media sites by spamming them with automated chatbot political campaigns.
In credit, Nay is not blind to the potential dangers of utilizing technology in this fashion.
As the research admits, “there may be a slow creep of less and less human oversight over automated assessments of policy ideas and the written communication to regulatory agencies and Congressional staffers. The core question raised is where to draw the line between human-driven and AI-driven policy influence.”
As AI chatbots continue to improve the question of where we draw the line may become fairly moot. Once artificial intelligence and human-led responses become indistinguishable, how can a line be drawn at all?
That’s a problem that currently holds no solution. When legislators don’t know what is composed by humans and what is written by AI, there is simply no practical way that humans can keep chatbots out of the political process.
and then