AI-Generated Comments Threaten Public Discourse, Raising Concerns for Regulatory Agencies

The rise of artificial intelligence is presenting a new challenge to government agencies relying on public feedback: distinguishing genuine citizen input from AI-generated comments. A recent case involving the South Coast Air Quality Management District (South Coast AQMD) highlights the potential for manipulation and the difficulty of ensuring authentic public participation.

The South Coast AQMD Case

Last year, the South Coast AQMD proposed a rule to incentivize the adoption of heat pumps over gas heaters. The agency received an unusually high volume of opposition – over 20,000 comments – far exceeding typical responses. This surge raised immediate concerns about the authenticity of the submissions. “Due to the volume and nature of these submissions, South Coast AQMD had concerns about their authenticity,” stated Rainbow Yeung, an agency spokesperson .

Further investigation revealed irregularities. The agency contacted a sample of 172 commenters to verify their submissions. Of the five who responded, three stated they had no knowledge of the comments submitted in their names. A Sierra Club campaigner reached out to four additional commenters, who also confirmed they hadn’t sent any emails .

The Los Angeles Times reported that CiviClick, a company specializing in “AI-powered advocacy tools,” orchestrated the opposition campaign on behalf of a public affairs consultant linked to the gas industry . CiviClick denies using AI to fabricate automated messages or submitting comments without consent.

The South Coast AQMD is continuing its investigation and exploring more robust methods for verifying commenter authenticity. The executive director indicated the team is considering “aggressive” sampling techniques, acknowledging the difficulty of drawing definitive conclusions from the initial limited response.

A Recurring Problem

The issue of fake comments is not new. In 2017, the Federal Communications Commission (FCC) received 22 million comments during the net neutrality debate, with approximately 18 million later identified as fraudulent . That instance involved a college student and a significant number of submissions originating from Russian email addresses. New York Attorney General Letitia James subsequently fined “lead generator” companies for impersonating millions of individuals during the comment submission process.

AI’s Role in Amplifying the Problem

While fake comments existed before AI, the technology has the potential to significantly amplify the problem. AI can now generate realistic-sounding text, making it easier to create and submit large volumes of deceptive comments. CiviClick maintains that its platform uses AI to help individuals personalize their comments by asking questions and tailoring email content. The company also uses AI to predict response rates to campaigns.

Implications for Public Discourse

The increasing sophistication of AI-generated content poses a fundamental threat to the integrity of public discourse and the ability of government agencies to make informed decisions based on genuine public input. As AI becomes more adept at impersonating humans, determining the authenticity of comments will become increasingly challenging, potentially undermining the democratic process.