Almost all of this year's big AI news has been about advancements in the technology. danger causing and speculation about how quickly technology will grow beyond the point of human control. But by 2024, governments have made significant inroads into regulating algorithmic systems. Here's a breakdown of the most important AI laws and regulatory efforts of the past year at the state, federal, and international levels.
state
US state lawmakers lead AI regulations by 2024, recommending Hundreds of bills—Some have small goals, such as creating a committee of education. While some may subject AI developers to serious civil liability. In cases where their creations cause serious damage to society. Most bills do not pass. But many states have enacted meaningful legislation that can serve as a model for other states or Congresses (assuming Congress will start working again).
If AI overwhelms social media before the election Politicians in both parties are behind more anti-Deepfake legislation. 20 states There is now a ban on deceptive AI-generated political ads in the weeks leading up to an election. Draft law aimed at regulating AI-generated pornography, especially images of minors It also has strong bipartisan support in states such as Alabama, California, Indiana, North Carolina. and South Dakota
Unsurprisingly, some of the most ambitious AI proposals have come from California. Because it is the backyard of the technology industry. One high-profile bill could force AI developers to take security precautions. and making companies liable for serious damages caused by their systems. The bill passed both legislative chambers amid intense lobbying efforts. But that's how it is. In the end it was vetoed. By Governor Gavin Newsom
However, Newsom has signed more than a dozen. Other bills It aims to expose less, but more immediate harm from AI. A new California law requires health insurance companies to ensure that the AI systems they use to determine coverage are fair and equitable. Another would require general AI developers to create tools that label content as AI-generated, and two bills would prohibit the distribution of AI-generated likenesses of dead people without their consent. in advance and requires that the agreement for AI-generated similarity of living people must clearly state how the content will be used.
Colorado passing average First time in US law Requires companies that develop and use AI systems to take reasonable steps to ensure that the tools are non-discriminatory. Consumer advocates call the law a Important basics– Similar bills are likely to be hotly debated in other states in 2025.
And I give the middle finger to both our future robot overlords and Utah Earth. Enact a law that prohibits government agencies from granting legal rights to artificial intelligence. Inanimate objects, bodies of water, atmospheric gases, weather conditions, plants, and other non-human objects.
central government
Congress has talked a lot about AI in 2024, and the House of Representatives closed out the year with its release. Both sides' reports, 273 pages Summary of guiding principles and recommendations for future governance. But when it comes to passing laws Federal lawmakers have done little to nothing.
Central government agencies, on the other hand, are Busy all year round It strives to meet the goals set forth in President Joe Biden's 2023 AI Executive Order and multiple regulatory agencies, most notably the Federal Trade Commission and the Department of Justice. It has cracked down on misleading and dangerous AI systems.
Agencies working to implement AI executive orders aren't particularly sexy or headline-grabbing. Rather, it lays an important foundation for the future governance of public and private AI systems. For example, federal agencies have embarked on a hiring spree for AI talent and created standard for responsible model development and hazard mitigation
And in an important step toward increasing public understanding of how government uses AI, the Office of Management and Budget has called on (most) joint agencies to release data. Important information about the AI systems they use that may affect people's rights and safety
FTC Enforcement Aspects Implementation of AI Target companies that use AI in deceptive ways, such as writing fake reviews or giving legal advice, etc. approved AI Gun Censor Company Evolv for Making Misleading Claims About What Their Products Can Do agencies as well decide An investigation into facial recognition companies IntelliVision which has been accused of falsely claiming its technology is free from racial and gender bias; and prohibit Drugstore chain Rite Aid from using facial recognition five years after investigation determined the company was using the tool to discriminate against shoppers.
Meanwhile, the DOJ has joined state attorneys general in suing real estate software companies. RealPage's large algorithmic pricing plans That has caused rents to increase throughout the country. It has also won several antitrust cases against Google, including one involving the company. Internet search monopoly This could dramatically change the balance of power in the growing AI search industry.
Around the world
in august EU AI Act effective– The law, which is already serving as a model for other jurisdictions, would require AI systems that perform high-risk functions, such as assisting with employment or medical decision-making. Risk must be minimized and certain standards regarding data quality, training and human supervision must be met. It also prohibits the use of other AI systems, such as algorithms that can be used to determine the social score of a country's residents. It is then used to deny rights and privileges.
in September China has issued AI security governance principles. frame– Like similar frameworks published by the US National Institute of Standards and Technology. This framework is not binding. Instead, it creates a common set of standards for AI developers to follow when identifying and mitigating risks in their systems.
One of the most interesting AI policies law Coming from Brazil, in late 2024 the country's Senate passed a comprehensive AI safety bill, facing a challenging road ahead. But if passed, it would create an unprecedented set of protections for different types of copyrighted content. commonly used to train creative AI systems Developers are required to disclose what copyrighted content is included in their training data. And creators will have the power to ban their use for training AI systems or negotiate compensation agreements that will be based, in part, on the size of the AI, the developer and how the materials are used.
The same is true of the EU AI Act. Proposed Brazilian law would also require high-risk AI systems to adhere to certain safety protocols.