Regulation of artificial intelligence

Artificial intelligence, or AI, is a type of computer that can learn and make choices like a person. Because AI is used more and more, many groups are making rules to keep it safe and fair for everyone. These rules help protect people and stop problems.

Since 2016, many groups have made ideas about the right way to use AI. Big companies that use AI also need to follow these ideas and be responsible.

In 2024, the European Union made special laws called the AI Act. This helps all countries in the European Union use the same rules. Groups like the IEEE and the OECD also help make plans for AI, even though they cannot make laws. Everyone wants AI to help society and be good for all people. Rules for AI are still being made, and this area is changing fast as AI gets better and more common. Artificial intelligence can change many parts of our lives, and good rules are important to help us all. AI ethics and making trustworthy AI are important parts of this work. The AI Act is one way countries are working together to make these rules.

Background

According to Stanford University, the number of laws and rules about artificial intelligence (AI) has grown quickly. Since 2023, mentions of AI in laws went up in many countries. In 2024, U.S. government agencies made many new rules about AI.

Some important leaders in AI technology have asked for rules to be made soon. People in different countries think differently about AI. In 2023, the United Kingdom began holding meetings about AI safety. More meetings happened in other countries later.

Perspectives

The regulation of artificial intelligence means creating rules and laws to guide how AI is used. Governments and experts want to make sure AI is safe, fair, and helpful. They worry about how fast AI changes and how to keep it safe while still sharing ideas.

Some experts think strict laws are hard because AI changes quickly. Others think flexible guidelines might work better. Some ideas suggest using special rules for certain AI systems to make sure they follow ethical practices. Young groups interested in AI have asked for stronger rules and teamwork between governments and companies.

Many agree on basic ideas, like keeping personal information private and making sure AI is safe and clear. In 2021, important leaders called for a special group to help guide AI rules. In 2025, the UK and US chose not to sign an international agreement on AI, saying it wasn’t clear enough.

Regulation can help make sure AI stays helpful in the long run. Some ideas include watching over AI research and focusing on safety.

Global guidance

In 2018, Canada and France suggested creating a global group to help guide how we use artificial intelligence. This group, called the Global Partnership on AI, started in 2020 with 15 countries. It aims to make sure AI is developed safely and fairly.

Many groups and countries have created guidelines to help manage AI. For example, the OECD and the G20 created their own AI principles in 2019. The United Nations has also started working on global standards for AI ethics. In 2023, the Group of Seven nations agreed on guiding principles for responsibly developing and using advanced AI systems.

Researchers and leaders are working together to create rules that protect human rights while allowing AI to advance for good.

Regional and national regulation

The rules and plans for managing artificial intelligence (AI) are changing around the world, in places like the European Union and Russia. Since early 2016, many countries and groups have started making plans and rules about AI. These cover many topics, like how to manage AI and how to use it in jobs, research, and building things.

Different countries handle AI in their own ways. For example, the United States focuses on letting businesses lead, China uses government control, and the European Union emphasizes protecting people’s rights.

The African Union is working on rules to protect people’s rights with AI, based on ideas like Ubuntu, which means caring for others.

In Australia, groups like the Australian Computer Society and others have asked the government to make a national plan for AI. In August 2024, Australia introduced a set of safety rules for AI, though specific laws are still pending. Existing laws, like the Privacy Act, also apply to AI.

Brazil passed a law in 2021 to guide AI use, focusing on ethics and human rights, but it faced criticism for being too vague. A new law in May 2023 calls for testing AI systems before they are used.

Canada has a national AI strategy with funding to support research and training. In November 2022, Canada introduced a law package to protect privacy and data, including the Artificial Intelligence and Data Act. In November 2024, Canada also announced a new institute to study AI safety.

China has been active in AI rules, with plans from 2017 and new rules in 2023 for AI services.

Colombia has several plans and guidelines to manage AI, though no specific laws yet.

The Council of Europe, which includes many European countries, began working on a treaty for AI in 2022. The treaty, adopted in May 2024, aims to protect human rights, democracy, and the rule of law in AI use.

The Czech Republic updated its national AI strategy in 2024 to include better laws and ethical rules.

The European Union has been a leader in AI rules, with the AI Act passed in May 2024. This law focuses on high-risk AI uses, like healthcare, and requires rules for transparency and human oversight.

Finland is preparing to follow the EU’s AI Act with its own national rules.

Germany released a detailed plan for AI standards in 2020 and updated it in 2022, covering many areas like safety and ethics.

Israel released its first national AI policy in December 2023, focusing on ethics and risk-based rules.

Italy approved rules in October 2023 for AI in healthcare, requiring transparency and human oversight.

Morocco is working on creating a national AI agency to manage AI use and increase awareness.

New Zealand currently uses existing laws to manage AI, with updates to the Privacy Act in 2020.

In the Philippines, a bill was proposed in 2023 to create an AI authority to oversee AI development and prevent crimes using AI.

Spain approved an AI research strategy in 2018.

Switzerland plans to adopt the Council of Europe’s AI treaty by February 2025 and will prepare laws by the end of 2026.

The United Kingdom released a plan for AI rules in March 2023, focusing on safety and innovation. In November 2023, the UK hosted a summit on AI safety and created an AI Safety Institute.

The United States has had discussions on AI rules, including an executive order in October 2023 on AI safety.

Regulation of fully autonomous weapons

Main article: Lethal autonomous weapon

Countries have talked about rules for weapons that can act on their own since 2014. These talks help make sure the weapons follow laws that protect people during conflicts. In 2018, experts agreed on some basic ideas about these weapons.

In 2016, China was the first big country to say that current laws might not be enough for these weapons. Some groups, like the Campaign to Stop Killer Robots, have asked for a worldwide stop to making and using these weapons. The United States thinks existing laws are enough to control them and says it does not currently have such weapons.