Artificial intelligence tools have many benefits in the fields of tax policymaking and lobbying, but the risks are real, too, according to a panel of tax experts at the recent ABA Section on Taxation’s May Tax Meeting in Washington, D.C.
May 20, 2024 Tax Law
The challenges of AI in tax policymaking
At the program “Making Tax Policy in the Era of Artificial Intelligence — Can It Survive the Robots?” panelists discussed AI’s impact on government services and the challenges confronting tax policymakers.
“AI can be a powerful tool for the public good. But there can be risks” that federal agencies are mandated to manage, said Varun Jain, senior counselor at the Office of Information and Regulatory Affairs in the Office of Management and Budget.
AI can be used to modernize regulatory review, said Mac Norman, an associate with Todd Welty PC, who focuses on tax litigation matters. It is useful for document review, tax compliance, discovery and database management, saving firms and clients time and money. AI also can be used for congressional lobbying, data and predictive analysis and monitoring changes in legislation and regulations in almost real time.
AI can make it easier for attorneys to draft letters to federal agencies for their clients on proposed regulations and can assist individuals in writing more thoughtful comments, which can number in the millions and help shape regulations, Jain said.
Yet humans still need to be involved to ensure AI summarization of comments is accurate, Jain added.
For AI use in the government, proper oversight is essential, said moderator Jeremiah Coder, deputy director of international tax policy at Revenue Jersey. Quality assurance and bias in AI systems are things to look out for.
Adrienne Griffin, a branch chief at the IRS Office of Chief Counsel, said confidentiality of taxpayer data is a major concern for the IRS, which is a reason the agency doesn’t use off-the-shelf AI systems. However, “the possibility of an in-house system is intriguing,” she said.
Alexandra Minkovich, a partner at Baker McKenzie, also cautioned about privacy concerns. She said attorneys need to be cognizant of what they share with AI tools because of ethics guidelines. Also, attorneys need to be very specific about the data shared with the tool to get the desired outcome in court cases. “It’s not just the capability of the AI tool, but the capability of the person using the AI tool,” she said.
When it comes to ethics and artificial intelligence, more questions exist than answers, Griffin said. She, too, emphasized the need to recognize inherent bias imbedded in AI software and tools and to ensure fairness when employing them.
“Act with commitment and dedication to your client’s interests,” Griffin said. And don’t use AI as a shortcut. “The human touch is still needed.”
Related links:
- ABA Task Force on Law and Artificial Intelligence
- GPSolo Magazine: “An introduction to artificial intelligence for lawyers”
- Legal Rebels podcast: How generative AI is already changing contract review
- ABA Journal: