Law Library of Congress Report on Safety and Security of Artificial Intelligence Systems

The Law Library of Congress in Washington, D.C. has published a new report on the Safety and security of artificial intelligence systems:

"This report surveys the safety and security of artificial intelligence systems (AI systems) in five selected jurisdictions, namely Australia, Canada, New Zealand, the United Kingdom (UK), and the European Union (EU)."

"The use of AI has increased exponentially and is permeating every aspect of our lives, from personal to professional. While it can be used in many positive ways to solve global challenges, there are also security risks to be considered, such as fundamental rights infringements, personal data security, and harmful uses. The European Union Agency for Cybersecurity (ENISA) has identified three dimensions to the relationship between cybersecurity and AI. First, there is the cybersecurity of AI, meaning a lack of robustness and the vulnerabilities of AI models and algorithms. Second, AI can also support cybersecurity when it is used as a tool or means to create advanced cybersecurity, such as by developing more effective security controls and by facilitating the efforts of law enforcement and other public authorities to respond to cybercrime. Lastly, there is the malicious use of AI, meaning when AI is used in a harmful, malicious, or adversarial way to create more sophisticated types of attacks. This report focuses on the first dimension."

"In order to ensure that AI systems are used to benefit society, jurisdictions around the world are looking into ways to regulate AI. Whereas the EU intends to adopt its legislative proposal for a specific Artificial Intelligence Act (draft AI Act) by the end of 2023 and the Canadian government introduced an Artificial Intelligence and Data Act (AIDA) in June 2022, other surveyed jurisdictions have not yet advanced similar legislation. Both Australia and New Zealand do not currently have laws or proposed specific laws related to AI. However, a May 2023 discussion paper published by the Australian government seeks public feedback on possible policy and regulatory responses to AI, including the adoption of a risk-based approach similar to the EU draft AI Act. Likewise, the UK has not passed AI-specific legislation and does not intend to do so currently, but might at a later stage 'enhance regulatory powers, ensure regulatory coordination, or create new institutional architecture.' Currently, there are 18 legal frameworks containing over 50 pieces of legislation that touch upon AI in the UK."

 The Law Library of Congress is the world’s largest law library, with a collection of over 2 and a half million volumes from all ages of history and virtually every jurisdiction in the world.

Over the years, it has published dozens of comparative law reports which are a treasure trove for legal research on a huge variety of issues.