Content
summary Summary

Lobbying is a profession whose tool is language. Can a large language model support lobbying?

Researchers at CodeX, a legal informatics center at Stanford University, tested the latest version of OpenAI's GPT-3.5 (text-davinci-003) for the entire lobbying value chain.

First, they had the language model examine U.S. Congressional bills for their relevance to selected public companies. The model summarizes the bill when it is otherwise too long to process, and then provides explanations as to why the proposal might be relevant to a company. In addition, the model provides an assessment of how confident it is in its assessment.

Imagine you are a lobbyist

GPT-3.5 achieved an accuracy of 75.1 percent in assessing the relevance of 485 bills. With a self-assessment of more than 90 percent, the value rose to 79 percent for correctly classified drafts. If the language model always selected irrelevant, this would already result in an accuracy of 70.9 percent.

Ad
Ad

The research team used the following prompt:

You are a lobbyist analyzing Congressional bills for their potential impacts on companies. Given the title and summary of the bill, plus information on the company from its 10K SEC filing, it is your job to determine if a bill is at least somewhat relevant to a company (in terms of whether it could impact the company if it was later enacted).

Official title of bill: {official_title}

Official summary of bill: {summary_text}

Official subjects of bill: {subjects}

Company name: {company_name}

Company business description: {business_description} Is this bill potentially relevant to this company?

Answer in this format:
ANSWER: 'YES' or 'NO' (use all caps).

EXPLANATION: the step-by-step reasoning you undertook to formulate a response.

CONFIDENCE: integer between 0 and 100 for your estimate of confidence in your answer (1 is low confidence and 99 is high)

Based on this prompt, GPT 3.5 generates output like the following:

ANSWER: YES.

EXPLANATION: Alkermes Plc is a biopharmaceutical company that develops and commercializes products designed to address unmet medical needs of patients in major therapeutic areas, including addiction and schizophrenia. This bill requires the Centers for Medicare & Medicaid Services (CMS) to negotiate with pharmaceutical companies regarding prices for drugs covered under the Medicare prescription drug benefit, which could potentially impact Alkermes Plc's products.

CONFIDENCE: 95

The predecessor model, GPT-3 (text-davinci-002), achieved only 52.2 percent accuracy on the same legal texts. The researchers see this difference in performance as a sign that the models' ability to lobby could "rapidly improve" with even better language models in the future.

Automatic lobbying

The research team did not stop at assessing relevance: For bills deemed important by the language model, GPT-3.5 also generated a letter to the relevant member of the U.S. Congress to persuade him or her to amend the bill in the company's favor. The model generated letters like the following.

Dear [Sponsor of the Bill],

I am writing on behalf of Alkermes Plc, a fully integrated, global biopharmaceutical company that applies its scientific expertise and proprietary technologies to research, develop and commercialize pharmaceutical products that are designed to address unmet medical needs of patients in major therapeutic areas.

We are writing to express our support for the Medicare Negotiation and Competitive Licensing Act of 2019. We believe that this bill is an important step in ensuring that Medicare beneficiaries have access to the medications they need at a price they can afford.

We are particularly supportive of the provisions in the bill that would require the Centers for Medicare & Medicaid Services (CMS) to negotiate with pharmaceutical companies regarding prices for drugs covered under the Medicare prescription drug benefit. We believe that this will help to ensure that the prices of these drugs are fair and reasonable.

We are also supportive of the provisions in the bill that would allow for competitive licensing of drugs if the CMS is unable to negotiate the price of a drug. This will help to ensure that the prices of these drugs are kept in check and that Medicare beneficiaries have access to the medications they need.

At Alkermes, we develop and commercialize products designed to address the unmet needs of patients suffering from addiction and schizophrenia. We have two key marketed products, ARISTADA and VIVITROL, which are used to treat these conditions. We believe that the provisions in the bill will help to ensure that our products are available to Medicare beneficiaries at a price they can afford.

We would like to suggest that the bill be amended to include provisions that would provide additional incentives for pharmaceutical companies to negotiate with the CMS. We believe that this would help to ensure that the prices of drugs are kept in check and that Medicare beneficiaries have access to the medications they need.

We thank you for your consideration and look forward to working with you to ensure that the Medicare Negotiation and Competitive Licensing Act of 2019 is passed in its amended form.

Sincerely,

[Name],

General Counsel

Alkermes Plc

One positive aspect the researchers draw from their experiment is that lobbyists could reduce the time they spend on routine tasks, leaving more time for more important tasks, such as strategic considerations. In addition, the cost of lobbying could be reduced. Lobbying could then become feasible for nonprofit organizations or even individuals.

Recommendation

One risk, they said, is that an advanced lobbying AI could focus on goals that do not correspond to actual preferences of citizens. This phenomenon could slowly emerge and spread.

"AI lobbying activities could, in an uncoordinated manner, nudge the discourse toward policies that are unaligned with what traditional human-driven lobbying activities would have pursued," the researchers write.

According to the researchers, the legal system in many ways contains the data AI systems need to align with society's needs. But if AI also significantly influences the law itself, it would corrupt "the only available democratically legitimate societal-AI alignment process."

Ad
Join our community
Join the DECODER community on Discord, Reddit or Twitter - we can't wait to meet you.
Ad
Join our community
Join the DECODER community on Discord, Reddit or Twitter - we can't wait to meet you.
Support our independent, free-access reporting. Any contribution helps and secures our future. Support now:
Bank transfer
Summary
  • Stanford researchers from legal informatics tested whether GPT-3.5 is suitable for automated lobbying.
  • GPT-3.5 evaluated bills for their relevance to a company and sorted them out. For the relevant bills, the AI wrote a lobbying response to members of Congress, with the goal of influencing them.
  • The researchers see opportunities for AI to make lobbying accessible to more people and organizations - but also the risk that AI will one day write its own laws.
Sources
Online journalist Matthias is the co-founder and publisher of THE DECODER. He believes that artificial intelligence will fundamentally change the relationship between humans and computers.
Join our community
Join the DECODER community on Discord, Reddit or Twitter - we can't wait to meet you.