The Biden Executive Order on AI: key takeaways
On 30 October, the White House issued an Executive Order on artificial intelligence ("AI"), encouraging the "safe, secure and trustworthy development and use of AI" (the "Order"). The Order also sets out disclosure requirements and industry-wide obligations for AI systems, which are widely defined to include any data system, software, hardware, application, tool or utility that operates in whole or in part using AI.
The Order was a clear signal from the US ahead of the UK's AI Safety Summit, which took place on 1 and 2 November 2023 that the US will continue to promote the regulation of AI, both internally and internationally, to ensure "it protects the public from potential harm and ensures that everyone is able to enjoy its benefits”. The Order focuses on the following key themes:
New Standards for AI Safety and Security
The Order directs developers to take various measures to protect American citizens from the potential risks of AI systems. For example, ensuring sufficient background testing of AI systems is carried out before they are made public. The requirements for this are twofold. Firstly, the National Institute of Standards and Technology will set standards for rigorous testing to ensure safety before public release. These will be applied by the Department of Homeland Security and the Departments of Energy and Homeland Security with particular scrutiny on chemical, radiological and nuclear risks and will be supplemented by further standards and best practices for detecting and clearly labelling AI-generated content. In addition, any companies developing AI systems which pose a risk to national security or public health must notify the US government when training any models and share the results of safety tests. The Order instructs the development of a National Security Memorandum to ensure US military and intelligence agencies use AI safely and effectively.
Protecting Americans’ Privacy
The Order recognises the power of AI to identify and exploit personal data. To help protect personal privacy the Order calls on Congress for bipartisan data privacy legislation to be passed. The Order goes on to prioritise federal support for 'privacy-preserving techniques, research and technologies' to preserve individuals' privacy. The Order also promises to strengthen guidelines for federal agencies in relation to privacy and the effectiveness of their current protection systems.
Wider Impact on Individuals
The Order contains some specific directions to protect individuals' rights and safety and mitigate the risks of discrimination and bias arising from AI systems. These include:
- Advancing Equity and Civil Rights - the Order directs that algorithmic discrimination will be addressed through training and assistance, including to landlords, contractors, benefit programs and in the criminal justice system.
- Standing Up for Consumers, Patients, and Students – in terms of healthcare and education, the Order says that it will advance the use of safe and secure AI practices in these sectors to mitigate the risks of misleading the public.
- Supporting Workers - the Order says that the US government will develop principles to reduce the risks of workplace surveillance, bias and job displacement and deploy these to employers. A report on 'AI’s potential labour-market impacts' will also be produced, analysing how increased federal support can be given to workers.
Promoting Innovation and Competition
The Order aims to promote fair and open competition by providing certain entrepreneurs access to assistance and resources to help with AI innovation. As part of this, a pilot of the National AI Research Resource will provide researchers with access to key AI data as well as grants for AI research. Existing authorities will also be asked to support the training of skilled immigrants to contribute to these areas within the US.
Advancing American Leadership Abroad
The Order recognises that global cooperation on AI systems is key. To that end, the Biden-Harris Administration have promised to expand lateral, multilateral, and multistakeholder frameworks to collaborate on the benefits and risks of AI. This will aim to promote the use of safe and responsible AI techniques contained in the Order at a global scale. This is to try and counter global challenges and enhance the implementation of safe AI practices across the world. As part of this, the Vice President spoke at the UK Government Summit on AI Safety on 1 and 2 November 2023.
Ensuring Responsible and Effective Government Use of AI
In the Order, the US Government recognised its own obligation to regulate the use of AI and help responsibly deploy AI within government systems. This would include the application of the other key themes discussed at a government level. The Order directed that help would be provided to government agencies with their obligations by issuing guidance on AI, increasing the hiring of AI professionals and acquiring AI products.
Our concluding thoughts
Practically speaking, the Order's effectiveness remains uncertain. The Order reflects the Biden-Harris Administration's roadmap for anticipated policy, and although a clear indication of Biden's intent, it relies on the effective coordination and enforcement of US agencies and multinational governments and companies (and so, has not yet imposed any new regulatory requirements perspective). As discussed in our deep dive, this is part of a developing patchwork of regulations around the world and within the US on the regulation of AI.