Monthly notes on digital technology and law

Just as society as a whole, businesses of all kinds are steeped in a process of deep digital transformation. The law is particularly important in this context, giving answers to questions such as, ‘who is liable if anything goes wrong?’, or, ‘what should regulatory oversight for any of these new digital technologies look like?’. Sensible laws, rules, and regulations can only evolve over time and legal guidance needs to evaluate many points of detail.

To support this process, we will share our monthly reviews of the developments that matter most at the intersection of digital technology and the law, looking at how developments are influencing and impacting one another – both across the technologies driving digitalization, and across jurisdictions.

For October, our monthly review may well have focused on the important announcements from the European Central Bank regarding the “digital euro” which, if launched, will be a sort of digital cash for the euro-zone. However, given how busy October has become regarding artificial intelligence (AI) policy and regulation, we would be remiss not to focus our attention here.

On 30 October 2023, the leaders of the Group of Seven (G7) officially announced their agreement on Guiding Principles and a Code of Conduct for advanced AI technology. These G7 Principles and their respective Code of Conduct provide guidance for organizations developing, deploying or using advanced AI systems, such as foundation models and generative AI. They establish commitments (among others) to (1) mitigate risks and identify vulnerabilities, (2) encourage responsible information sharing, (3) report incidents, (4) invest in cybersecurity, and (5) enable users to identify AI generated content.

It is expected that these G7 documents will have a significant practical effect on national legislation as well as on corporate policies of business organizations. Going forward, they will form a point of reference of any discussion on AI governance. Read more details on this important development in our HL Engage publication authored by Louise Crawford and Sam Krips.

On the same day, U.S. President Joe Biden signed a sweeping Executive Order on the Safe, Secure and Trustworthy Development and Use of AI. It is expected that this executive order will trigger the release of a wide array of regulations for AI applications as it directs a significant number of U.S. government agencies and other stakeholders to take action, including (among others) for:

• Developers of foundation models that pose a serious risk to national security, national economic security, or national public health and safety to (1) notify the federal government when training the model and (2) share the results of all red-team safety tests;
• The National Institute of Standards and Technology (NIST) to set standards for extensive red-team testing to ensure safety before public release of models;
• The Department of Commerce to develop guidance for content authentication and watermarking to clearly label AI-generated content;
• Various federal agencies to consider how to address privacy concerns regarding the development and deployment of AI models;
• The U.S. Patent and Trademark Office as well as the U.S. Copyright Office to issue recommendations regarding the AI related legal implications for patents and copyrights respectively;
• The Federal Trade Commission to consider competition risks associated with the concentration of control of AI technologies;
• The Labor Department to issue guidance on minimizing the risk of bias arising from the use of AI in the workplace.

Meanwhile, the European Union (EU)’s draft of the AI Act has gone into the final round of its legislative process with the trialogue negotiations between the EU Parliament, the EU Council and the EU Commission currently taking place. There are only three points yet to be determined:

• The exemptions from “prohibited AI applications” for the benefit of national law enforcement authorities – in particular with regard to the scope of real time facial recognition in public places;
• The specific regulation that may be useful for AI foundation models – as the current draft seems quite half-baked on this issue;
• The level of fines and other sanctions that shall apply in case of incompliance with the requirements of the AI Act.

We expect this legislative process to be concluded at some point later this year so that the EU's AI Act will be enacted soon thereafter.

As we kicked off November, the UK hosted the AI Safety Summit and released the Bletchley Declaration, which has 29 signatories, including China, the U.S., and the EU. Stay tuned for more on this from our team in London, available on the AI Hub soon.

We will be back in a month from now with an overview of this and the other relevant developments at the intersection of digital technology and law to come in November.

[View source.]