Late in 2023, the corporate world was abuzz with a fervent, almost utopian promise regarding the capabilities of Artificial Intelligence – particularly its potential to revolutionize environmental sustainability. Tech giants and academic institutions alike painted a future where AI would act as a powerful catalyst for positive ecological change, tackling some of humanity’s most pressing climate challenges. This narrative positioned AI as a benevolent force, an indispensable tool in the global fight against climate change, capable of optimizing resource consumption, predicting environmental shifts, and accelerating the development of green technologies. The widespread enthusiasm was palpable, fostering a sense of optimism that cutting-edge technology could indeed offer a path to a more sustainable planet.
From today’s vantage point, just a few years later, many of the corporate claims circulating back then sound remarkably audacious, bordering on the absurd. Google, a leading proponent of AI development, insisted in a public statement linked to COP28 that AI possessed the inherent potential to “mitigate 5-10 percent of global greenhouse gas emissions.” This bold assertion suggested AI could single-handedly offset a significant portion of humanity’s carbon footprint through applications such as smart energy grids, optimized logistics, and advanced climate modeling. Not to be outdone, Microsoft enthusiastically declared that the nascent technology would “accelerate the discovery and development of sustainability solutions,” envisioning AI as a supercharger for scientific breakthroughs in renewable energy, carbon capture, and waste reduction. Even the academic community contributed to this chorus of optimism. Jim Bellingham, a prominent figure from Johns Hopkins’ Institute for Assured Autonomy, claimed in March 2023 that “AI-powered robots” and sophisticated satellite systems would integrate into a comprehensive global network, actively working to “reduce the carbon that is released into the atmosphere.” These statements, made with considerable conviction, laid the groundwork for a public perception of AI as a powerful ally in environmental conservation.
However, the unfolding reality has starkly contradicted these lofty predictions, demonstrating almost the exact opposite trajectory. Far from being a net positive for the environment, the foundational infrastructure supporting AI has emerged as a significant and rapidly growing contributor to global carbon emissions. The insatiable electrical demands of modern data centers, which house the vast computational power required to train and run complex AI models, are supercharging our carbon footprint at an alarming rate. These sprawling facilities consume colossal amounts of energy, often equivalent to the power consumption of small cities or even entire nations, for both computation and the extensive cooling systems necessary to prevent overheating. Moreover, beyond its direct energy consumption, AI technology is now being actively deployed in ways that resist and undermine crucial climate regulations, creating a deeply troubling paradox where a supposedly green technology is used to obstruct environmental progress.
A particularly stark example of this counter-intuitive development emerged from new reporting by the *Los Angeles Times* in February 2026. Their investigation uncovered how an otherwise promising eco-friendly initiative in Southern California, aimed at gradually phasing out gas-powered appliances, was decisively defeated through a sophisticated campaign that weaponized a suite of AI software. The regulation, proposed by the South Coast Air Quality Management District (AQMD), represented a forward-thinking effort to improve regional air quality. Specifically, it sought to incrementally phase out household water heaters and furnaces that relied on natural gas, as part of a broader strategy to limit the emissions of smog-causing nitrogen oxide (NOx). NOx is a major component of ground-level ozone, which contributes to respiratory illnesses, cardiovascular problems, and environmental degradation, particularly in a highly populated and often smog-ridden region like Southern California. The AQMD’s proposal was a critical step towards cleaner air and a healthier future for millions of residents.
As the initiative proceeded through AQMD’s board in the summer of 2025, the agency found itself unexpectedly inundated with tens of thousands of emails, all voicing opposition to the proposed regulations. These communications, initially appearing to be organic expressions of public sentiment, overwhelmed the district’s capacity and created a powerful impression of widespread dissent. As the *LA Times* meticulously documented, this flood of seemingly angry public comments received a significant, orchestrated boost from CiviClick, an AI-powered platform that boldly bills itself as “Disruptive Digital Advocacy Software.” This tool promised to revolutionize grassroots campaigning by automating and amplifying public outreach, ostensibly giving a voice to constituents. However, its application in this context raised immediate questions about the authenticity and origin of the public’s “voice.”
According to records requests meticulously obtained and analyzed by investigative journalists, a staggering number – more than 20,000 public comments – directed at AQMD’s initiative were directly drummed up via the CiviClick platform. This sheer volume was unprecedented for the agency, which typically receives only a handful of comments per agenda item. While the precise extent to which AI was deployed in crafting each individual message within this campaign remains opaque, CiviClick openly boasts a comprehensive suite of solutions bundled under the enticing moniker of “Grassroots Advocacy Software.” These tools are designed to streamline and amplify public engagement, offering features such as auto-filled webforms that simplify the submission process for individuals, as well as functionalities for generating and sending video and photo messages directly to elected officials. Most crucially, and most relevant to this controversy, CiviClick prominently advertises an “AI powered automatic message generator.” This advanced feature comes equipped with “auto-randomizing” messaging, allowing for the creation of unique, yet thematically consistent, communications. It also promises “unlimited subject lines and message bodies,” ensuring that no two emails appear identical, thereby bypassing simple spam filters and creating a more convincing illusion of individual sentiment. Furthermore, its “smart targeting ability” allows campaigns to precisely identify and engage specific demographics or geographic areas, maximizing impact. These capabilities, when deployed en masse, create a formidable, automated advocacy machine.
In the face of such vast public opposition – or what appeared to be, at the time, an overwhelming tide of citizen outrage – the AQMD board ultimately buckled under the pressure. In a pivotal vote on June 6, 2025, the board voted 7-5 to reject the measure. This critical decision effectively shelved the proposal, which would have incrementally raised the cost of gas-powered appliances through small surcharges, thereby incentivizing the transition to cleaner, electric alternatives over time. The defeat represented a significant setback for environmental advocates and a victory for the fossil fuel industry, halting progress towards cleaner air in a region desperately in need of it.
The mastermind behind this sophisticated email campaign, Matt Klink, was unequivocal in his assessment of CiviClick’s impact. He went so far as to claim that the AI platform made “the ultimate difference” in successfully shooting down the proposed environmental safeguards. In a sponsored interview with the political consulting publication *Campaigns and Elections*, Klink elaborated on the campaign’s efficacy: “We did two separate rounds of outreach, and they met the targets in both rounds early,” he stated, highlighting the platform’s efficiency in generating the desired volume of opposition. He underscored the unprecedented nature of the deluge, noting, “AQMD staff are not used to getting tens of thousands of emails so it [CiviClick] made a massive difference in turning the tide.” Klink’s candid remarks illuminated the strategic intent behind employing such a powerful, automated advocacy tool to sway public policy.
Despite Klink’s assertion that CiviClick was merely an advanced email tool designed to connect voters with AQMD staff en masse, an investigation by the *LA Times* quickly uncovered disturbing evidence to the contrary. The newspaper’s diligent reporting revealed that at least three individuals, when contacted afterward, explicitly stated they were entirely unaware that CiviClick had sent comments to AQMD on their behalf. This finding strongly suggested a deceptive practice, where names and contact information were potentially used without explicit consent to generate opposition, undermining the very notion of genuine grassroots advocacy. It exposed the astroturfing nature of the campaign – creating the illusion of widespread public support or opposition where none genuinely existed.
Sources within AQMD, speaking anonymously to the *LA Times*, confirmed that the sheer onslaught of emails and comments definitely had a profound impact on the board’s decision-making process. The unprecedented volume created an impression of overwhelming public resistance that was difficult for board members to ignore. This manufactured public outcry, whether genuine or not, played directly into the hands of the gas industry. For its part, the industry had already been actively waging a legal battle against the environmental agency since December 2024, signaling its strong opposition to any regulations that threatened its market share. The AI-powered email campaign provided an additional, potent weapon in their arsenal, giving them a “public” voice that might not have otherwise materialized. As one staffer candidly told the *LA Times*, the usual amount of public comments received for any given agenda item could typically be counted on one hand, emphasizing the extraordinary and suspicious nature of the CiviClick-generated deluge.
Notably, Matt Klink conspicuously declined to disclose to the *LA Times* who had funded the CiviClick campaign, maintaining a veil of secrecy around the financial backing of this influential operation. His public affairs company, Klink Campaigns, is a partner at California Strategies, which the paper identified as one of the largest and most influential lobbying firms in the state. This connection immediately raised red flags, pointing to the deep ties between the campaign and powerful corporate interests. California Strategies’ client roster includes significant players such as corporate landlord groups, energy conglomerates, and prominently, the Fortune 500 energy company Sempra. Sempra, in turn, is the parent company of the massive Southern California Gas Company (SoCalGas), a direct beneficiary of the defeat of the AQMD’s regulation. The refusal to disclose funding, coupled with these powerful corporate connections, painted a clear picture of a well-resourced, industry-backed effort to undermine environmental policy.
While astroturfing – the practice of masking the sponsors of a message or organization to make it appear as though it originates from grassroots participants – is nothing new in US politics, the deployment of AI-powered astroturfing marks an ominous and unprecedented escalation. This development signals a horrifying turn for genuine civic engagement, as the ability to discern authentic public sentiment from algorithmically generated noise becomes increasingly difficult. Furthermore, it opens up a tremendous “can of worms” for political campaigning across the United States, threatening to fundamentally alter the landscape of democratic discourse. As long as AI can fabricate thousands of seemingly unique constituent messages on demand, corporate and special interests no longer need to maintain the pretense of widespread public support or opposition. Instead, they can simply rent it, effectively commodifying and manipulating the very mechanisms of democracy for their own gain. This trend poses a profound threat to the integrity of public policy and the legitimacy of citizen participation.