States Pioneering Diverse Approaches to Artificial Intelligence Governance ~ The Lawyers Clock

States Grapple with AI Legislation: A 2025 Landscape

The rapid expansion of artificial intelligence across all sectors in the United States has spurred a wave of legislative proposals. From deepfakes to automated decision-making, state lawmakers are trying to figure a path through the tricky parts of AI regulation. Some efforts are geared toward protecting citizens from AI misuse, while others are designed to encourage innovation. In this opinion editorial, we take a closer look at how the states are managing their way through AI’s evolving legal environment as of 2025.

This discussion digs into the state-level responses to AI regulation, offering a panoramic view of recent proposals that reflect divergent political approaches. While the focus is largely on shielding the public from potential harm, the underlying tension between protection and innovation remains clear. As states pursue their own strategies, the federal government’s looming role adds another layer of complexity to an already nerve-racking issue.

Understanding the State-by-State Approach to AI Oversight

As of early 2025, data shows that 34 states have embarked on studies related to artificial intelligence, with 24 states creating special groups to study the technology and another 10 forming standing committees. In total, lawmakers have introduced nearly 260 AI-related measures during the legislative session, with only a fraction having turned into law. This diverse array of proposals represents an evolving, yet immature, market where the focus is on protecting citizens.

One common observation by policymakers is the frequent use of words like “prohibit” and “disclosure” in these bills. Such language underscores a primary objective: shielding citizens from potential AI overreach rather than embracing the technology for enhanced services. The directives range from measures aimed at nonconsensual imagery to those regulating government use and automated decision-making systems.

This state-level activity comes against a backdrop of national competition. While 54 countries have already published their national AI plans — some emphasizing defense-related initiatives, others focusing on societal betterment — the U.S. also sees a need to build human talent and collaborations. Efforts like the “Computer Science For All Act” and the Partnership for Global Inclusivity on AI (PGIAI) illustrate how federal initiatives are striving to create a more robust ecosystem for AI development. Yet, the states continue to drive forward with their unique legislation in parallel.

Deep Dives into AI-Related Legislative Areas

The legislative proposals addressing AI in 2025 can be broadly categorized into a few key areas. In this section, we take a closer look at major topics and discuss how states are sorting out policy in each one.

Protecting Citizens from Nonconsensual Imagery and Related Harms

One major area of focus is nonconsensual intimate imagery (NCII) and child sexual abuse material (CSAM). Several states have introduced bills aimed at curbing the distribution of synthetic or deepfake content created without consent. The aim is to hold online platforms accountable and ensure that victims of such digital abuses receive protection.

For instance, Maryland and Mississippi have introduced bills that seek to impose hefty penalties on those who knowingly disseminate synthetic NCII. Maryland’s TAKE IT DOWN Act, for example, mandates that designated online platforms establish processes for quick removal of harmful content. Meanwhile, New Mexico’s proposal targets the nonconsensual public dissemination of synthetic images tied to identifiable individuals. Although many of these bills have stalled in committee, their introduction indicates a strong legislative focus on this sensitive topic.

The approaches in these states highlight a broader trend: legislatures are trying to find their way through the tangled issues presented by new technologies, ensuring that the law keeps up with both traditional and digital forms of abuse. Some of the notable measures include:

Mandated removal processes for harmful synthetic content.
Hefty penalties to deter malicious use of deepfake technology.
Requiring platforms to have systems in place for rapid content elimination.

Election Integrity and AI: Keeping Campaigns Transparent

As political campaigns face pressures from AI-generated content, election-related legislation has become another battleground. State lawmakers in 2025 have been particularly concerned with ensuring that AI is not used to mislead or manipulate voters. Bills have been proposed that require candidates and political campaigns to disclose any use of synthetic content in advertisements.

For example, in New York, a proposed bill would force political communications that incorporate synthetic media to clearly state that AI was used. Similarly, Massachusetts has put forward a bill that seeks to prevent the creation and spread of deepfake videos or altered images intended to tarnish the reputation of political candidates. Although these measures are still finding their course in committee, they reflect a bipartisan priority to maintain transparent and fair elections.

Key components discussed in these proposals include:

Mandatory disclosures of AI involvement in political advertising.
Pre-election bans on misleading deepfake content.
Oversight measures to prevent AI-generated misinformation from impacting elections.

These proposals aim to address not just the overreach of AI in elections but also the nerve-wracking potential for digital manipulation at a time when trust in media and governance is especially fragile.

Transparency Regulations for Generative AI Systems

Another area of significant concern highlighted by state legislators is the transparency of generative AI. The rise of sophisticated AI-driven chatbots and other tools has sparked worries that consumers may unknowingly interact with machines rather than humans, leading to a series of legal and ethical debates.

Legislation in states like Hawaii and Massachusetts specifically addresses these issues by requiring that entities engaged in commercial activities clearly notify consumers when they are interacting with an AI system. Such bills typically propose that companies not only inform users but also establish protocols—like red team assessments—to check that embedded watermarks cannot be easily removed from AI-generated content.

This approach to transparency ensures that users are aware of who—or what—they are engaging with, avoiding any confusion or deception. The provisions outlined in these measures include:

Clear and conspicuous notifications about AI involvement in customer interactions.
Regular audits or red teaming of AI systems to check for watermark vulnerabilities.
Periodic reports to state authorities regarding the robustness of AI transparency measures.

By creating these safeguards, states hope to build consumer trust while striking a balance with innovation. The debates over whether these regulations hinder business innovation or promote responsible AI adoption are ongoing, offering plenty of fodder for future discussion.

Regulating Automated Decision-Making and High-Risk AI

The surge of interest in automated decision-making technology (ADMT) and high-risk AI applications has compelled lawmakers to draft proposals aimed at mitigating potential harms. The central idea behind these bills is the implementation of safeguards that ensure decisions affecting people’s lives are not driven solely by opaque algorithms.

A good example of such an initiative is Colorado’s AI Act, which focuses on algorithmic discrimination in consequential areas such as credit, employment, or public services. The Colorado model mandates transparency and accountability on the part of both developers and deployers of AI systems, setting a benchmark that other states have looked to emulate in their own proposals.

Legislative proposals in states including Georgia, Illinois, Iowa, and Maryland have attempted to introduce similar frameworks. Notable aspects include:

Requiring disclosure when AI is used as a substantial factor in decision-making.
Mandating transparency regarding algorithms and their potential biases.
Establishing a duty of care on businesses and developers to ensure fair outcomes in consequential decisions.

These measures are not without their twists and turns. While some proposals have made significant headway, others have stalled in committee. The debate remains heated: is stringent regulation a necessary safeguard, or does it risk stifling innovation in a sector already navigating a maze of confusing bits and complicated pieces?

Government Use of AI and the Role of Accountability Boards

Legislation around government use of artificial intelligence represents another vital frontier in the AI debate. States such as Georgia, Montana, and Nevada have taken markedly different paths in addressing the role of AI in administrative decision-making and public sector applications.

Georgia’s proposed “AI Accountability Act” suggests establishing a dedicated board to develop comprehensive plans for AI usage, including robust data privacy measures and clear outlines of human oversight responsibilities. In contrast, Montana has enacted legislation that restricts state and local government from relying too much on AI for decisions that require human judgment, underscoring the view that certain responsibilities should remain firmly in human hands.

Key provisions of these bills tend to follow similar themes:

Setting up oversight bodies or accountability boards to monitor AI use by government agencies.
Requiring the disclosure of AI system usage in official communications.
Mandating human review of critical decisions or recommendations influenced by AI.

These initiatives reflect an overarching goal: to prevent the state itself from becoming a source of harm through misapplied digital tools. Many lawmakers see these measures as essential steps to ensure that AI remains a tool for enhancing public service rather than an off-putting replacement for human judgment.

Workplace Protections in the Age of AI

Employment-related legislation is also riding high on the AI debate. With algorithms increasingly used in recruiting, hiring, and performance monitoring, workers and job applicants are calling for measures to ensure fair treatment and transparent practices. Several state proposals target the use of AI in employment decisions, seeking to protect against hidden biases and unobvious automated screenings.

For example, in Illinois and Pennsylvania, proposed bills would require employers to inform job candidates when their applications or interviews are processed by AI systems. Similar legislation in California aims to curb excessive workplace surveillance powered by AI, allowing employees to be aware of the data collected about them.

The main points emphasized in these proposals include:

Notifying applicants and employees when AI is involved in screening or decision-making processes.
Implementing restrictions on intrusive AI-based surveillance methods at the workplace.
Ensuring transparency in any automated processes that could affect employment status or conditions.

By focusing on these protections, state legislatures aim to carve out a safe space where the benefits of AI in streamlining operations do not come at the expense of worker rights. While the proposals remain under review in many states, the conversation has already sparked broader debates on the delicate balance between innovative hiring practices and the preservation of individual rights.

Health Care, AI, and Patient Protection Laws

The intersection of artificial intelligence and health care presents a particularly nerve-racking challenge for legislators. As AI tools are increasingly integrated into clinical decision-making, treatment planning, and patient care communications, lawmakers are concerned about the potential for misdiagnosis, breach of privacy, or even the deprofessionalization of therapy.

California, Illinois, and Indiana have introduced legislation aimed at safeguarding patients while still allowing for technological progress. California’s proposal, for instance, is carefully designed to protect consumers from AI systems that might imply unearned professional competence. Meanwhile, Illinois is considering a bill that would bar licensed health care professionals from relying on AI for making therapeutic decisions or formulating treatment plans without proper human oversight.

Other states, like Indiana, are proposing less expansive measures, focusing on disclosure requirements. These bills typically mandate that health care providers must inform patients if AI is involved in any decision-making process pertaining to their treatment. The recurring themes in these laws include:

Restrictions on AI-driven therapeutics without explicit human oversight.
Disclosure requirements so that patients know when AI contributed to a diagnosis or treatment plan.
Measures aimed at preventing AI from overshadowing the role of licensed professionals.

In the realm of health care, the stakes are exceptionally high. Legislators must figure a path through a minefield of potential issues while striving to maintain the integrity of patient care and trust in the medical profession.

The Political Dynamics of AI Legislation

One cannot discuss state-level AI legislation without considering the political dynamics that influence these proposals. The statistics from 2025 indicate that approximately two-thirds of the introduced bills were sponsored by Democrats, with Republicans contributing roughly one-third. Only a few bills found common ground between the parties, which illustrates the consistently politically charged environment surrounding AI regulation.

Democrats have generally pushed harder on establishing comprehensive AI governance frameworks, reflecting their broader tendency toward tighter tech regulations. Meanwhile, Republican-led states tend to advocate for lighter governmental interference, favoring measures that prioritize innovation while still curbing particularly harmful uses of AI. Among the politically sensitive topics, election-related deepfake bans and child pornography proposals have garnered bipartisan interest, even though more sweeping regulatory measures remain politically polarizing.

This divergence in legislative approach is demonstrated through a summary table of the major themes and their corresponding legislative metrics from 2025:

Legislative Category	Bills Introduced	Bills Signed into Law
Nonconsensual Imagery / CSAM	53	0
Elections	33	0
Generative AI Transparency	31	2
Automated Decision-Making / High-Risk AI	29	2
Government Use	22	4
Employment	13	6
Health Care	12	2

This table vividly showcases the fact that while the legislative spark is strong, many of these proposals remain in a state of flux, having yet to clear the often intimidating legislative hurdles. The political differences not only shape the content of the bills but also determine their pace and eventual impact.

The Federal Factor: A Looming Challenge

While states take the wheel in experimenting with AI regulation, there is growing concern about the potential interference from federal lawmakers. Recent maneuvers in the U.S. Senate — including discussions about a temporary moratorium on state-level AI laws — have added a new wrinkle to an already complicated debate. Although the proposed moratorium did not come to fruition, the fact that such measures were even raised signifies a federal interest in consolidating AI regulations.

Federal proposals and action plans, such as the AI Action Plan which tasks agencies like the FCC with monitoring the implications of state laws, introduce an overarching narrative that might eventually pre-empt state efforts. The uneasy balance between federal oversight and state autonomy creates a scenario where local initiatives might be either bolstered or stymied by national policies.

Key challenges arising from this interplay include:

The risk that federal standardization may override tailored state solutions.
Potential conflicts between state-driven protections and nationwide innovation benefits.
The possibility that a unified federal approach could dilute state-level experimentation and progress.

Many legal experts express concern about what might happen if federal authorities decide to impose a one-size-fits-all model on an issue that is inherently loaded with regional variations and political subtleties. At this juncture, it remains an open question whether federal interventions will hinder or harmonize the incremental progress being made at the state level.

Looking Ahead: Prospects and Policy Playbooks for State Legislatures

Despite the current uncertainties at the federal level, the states appear committed to pressing hard on AI legislation. The dynamic nature of these proposals suggests that state governments are using the current session as a testing ground for various regulatory frameworks. Whether it’s through comprehensive laws like Colorado’s AI Act or more targeted proposals addressing specific issues, lawmakers are actively engaged in sorting out the small distinctions between protection and progress.

As the debate continues, several essential strategies have emerged as potential best practices for states looking to craft robust AI legislation. These strategies include:

Adopting a Multi-Stakeholder Approach: Engaging tech companies, legal experts, and consumer advocates in the drafting process helps create policies that balance fairness with innovation.
Fostering Transparency and Accountability: Clear disclosure requirements and oversight mechanisms can build public trust while ensuring that AI's role in shaping public life is visible and accountable.
Implementing Layered Regulation: Combining broad legislative frameworks with issue-specific rules allows lawmakers to address both tangible harms (such as in NCII or election deepfakes) and the subtle parts of AI’s impact.
Encouraging Pilot Programs: Testing new measures in smaller jurisdictions or in limited scopes can provide useful feedback before rolling out statewide mandates.

These measures represent a playbook that other states can adapt, resizing and refining their regulations as the industry matures. The overarching goal is to ensure that AI serves the public interest without unnecessarily hampering technological innovation.

Furthermore, as states continue to refine these policies, it is crucial for them to engage in cross-state dialogues. Sharing lessons learned, successes, and setbacks can help build a more cohesive national framework—one that respects regional differences while establishing consistent standards for AI governance.

Conclusion: Balancing Protection and Progress

The evolving landscape of AI regulation at the state level in 2025 illustrates a vibrant, if sometimes chaotic, democratic experiment. Lawmakers are working diligently to protect citizens from the potentially harmful twists and turns of AI technology while still allowing room for the innovation that may drive economic growth and improve public services.

Protections focusing on nonconsensual imagery, election transparency, and workplace and health care safeguards demonstrate a clear commitment to shielding the public. However, these efforts are loaded with challenges—from the delicate balance of bipartisan political interests to the looming shadow of federal intervention.

As debates continue in legislative halls and committees, the path forward is sure to be both exciting and nerve-racking. State lawmakers are not merely drafting laws; they are piecing together a framework that may ultimately define how the nation navigates AI’s promising, yet complicated, future. The legal and societal impacts of these decisions will reverberate for years to come, demanding careful attention from all stakeholders.

In this dynamic environment, each state’s approach provides critical insights into the little details and broader policy directions needed to effectively harness AI for the public good. The balancing act between offering protections and promoting progress is a central theme of the day, and its resolution will likely shape the trajectory of American innovation in the digital age.

Ultimately, as the states continue to innovate and experiment with AI regulation, it falls upon all branches of government—state and federal alike—to work through the convoluted twists and turns of this emerging technological era. The forthcoming years will determine whether these policies can effectively safeguard citizens while nurturing an environment where technological advancements can shine, free from oppressive overregulation.

Through ongoing dialogue, mutual learning, and a willingness to adapt, the United States has the opportunity to craft a balanced, responsive legal framework that serves as a model both domestically and internationally. In the meantime, all eyes remain on state legislatures as they figure a path through one of the most critical regulatory challenges of our time.

Originally Post From https://www.brookings.edu/articles/how-different-states-are-approaching-ai/