The United States, European Union, United Kingdom, and several other countries have signed the first international agreement on artificial intelligence in Vilnius.
According to the legally binding agreement, countries are required to implement safeguards against any threats AI may pose to human rights, democracy, and the rule of law. The treaty, named the Framework Convention on Artificial Intelligence, was drafted by the Council of Europe.
According to the Council of Europe, the convention aims to fill any legal gaps that may arise due to rapid technological advancements. Recent developments in AI have triggered a global regulatory push to mitigate the potential drawbacks of the technology.
This means there is a range of regulations and agreements covering the technology, from the EU’s AI laws to last year’s Bletchley Declaration at the inaugural global AI safety summit, as well as a voluntary testing regime signed by numerous countries and companies at the same event.
The Vilnius Agreement represents an attempt to create a global framework for AI. The convention will come into force three months after ratification by five signatories, including at least three Council of Europe member states, and countries from around the world are eligible to sign the convention.
The treaty stipulates that AI systems must comply with a set of principles, including data protection, non-discrimination, secure development, and human dignity. As a result, governments are expected to introduce safeguards such as preventing AI-generated disinformation and ensuring systems are not trained on biased data, which could lead to flawed decisions in various situations, such as job applications or benefits.
The agreement also covers the use of AI by public authorities and the private sector. Any company or entity using relevant AI systems must assess their potential impact on human rights, democracy, and the rule of law and make that information available to the public.
People must have the ability to challenge decisions made by AI systems and appeal to authorities.
AI system users must also be informed that they are interacting with artificial intelligence, not a human being.