Eka’s Weekly Roundup (3 November 2023)

Issue 51 l UK and US updates on AI regulation

Nov 03, 2023

Two big developments in AI regulation this week: Joe Biden’s executive order on Monday, which lays out new standards for AI safety and security, and the UK’s AI Safety Summit hosted in Bletchley Park on Wednesday and Thursday. We summarise the takeaways below.

Updates on AI regulation 🗞️

What happened on Monday: Biden’s executive order

President Joe Biden's recent executive order on artificial intelligence (AI) has been hailed as a significant step forward in regulating the rapidly advancing technology. The order, spanning nearly 20,000 words, outlines the need for new federal standards focused on AI safety, security, and trustworthiness. We list some of the key points below (White House).

Standards

Requiring developers to share safety test results with the US government
Develop standards to help ensure the safety of AI systems
Protect against the risks of using AI to engineer dangerous biological materials
Protect citizens from AI-enabled fraud
Establish cybersecurity programs to develop AI tools to find and fix software vulnerabilities.

Equity and Civil Rights

Guidance to landlords, benefits programs, and contractors to avoid AI algorithms from discriminating.
Address algorithmic discrimination in training, technical assistance, and coordination in the justice system.

Consumers, Patients, Students

Responsible AI in the development of affordable drugs
Address AI’s potential to transform education

What happened on Tuesday: Lukewarm response to the EO

While the order creates various task forces and advisory committees, prompts reporting initiatives, and directs federal agencies to issue guidelines on AI within the next year, experts emphasized some limitations (Scientific American). The order does not specifically address the protection of biometric data privacy, leaving a crucial area unregulated. There are concerns about evaluating and mitigating AI bias and discriminatory algorithms, as well as gaps in addressing the government's use of AI in defense and intelligence applications.

Furthermore, technical feasibility poses challenges. For instance, the order mandates the Department of Commerce to identify best practices for labelling AI-generated content within eight months, but there is currently no robust method for accomplishing this task.

While experts acknowledge the order's significance, they emphasize that it is not a comprehensive solution. Legislative action is crucial for addressing all the complexities associated with AI, and the order itself acknowledges the importance of Congress passing data privacy legislation. Despite its impact, the policy is not expected to bring immediate changes to people's everyday interactions with AI technology. Nonetheless, it represents a significant stride toward regulating AI, filling a policy void that the United States has been grappling with amid the rapid evolution of this transformative technology.

What happened on Wednesday & Thursday: UK’s AI Safety Summit

The AI Safety Summit 2023, hosted by the UK Prime Minister at Bletchley Park, Buckinghamshire, brought together global governments, leading AI companies, experts, and civil society groups. The summit focused on addressing the risks associated with frontier AI development and discussing international coordination to mitigate these risks. Here are a few takeaways published by The Guardian.

Diplomatic Success: Prime Minister Rishi Sunak successfully convened global leaders, tech executives, and academics, leading to the signing of an international declaration involving over 25 countries and the EU to address AI risks. The summit was considered a diplomatic coup.
US Dominance in AI: The US demonstrated its influence by setting the AI agenda, as seen in the EO published earlier in the week. Vice President Kamala Harris announced the establishment of an AI safety institute, showcasing the US's political and commercial power in AI.
Elon Musk's Impact: Elon Musk's presence at the summit added attention and star power. Despite his relatively smaller AI venture, his warnings about AI's threat to humanity captured significant attention.
Immediate Concerns: While the tech community remains divided on the existential risk of AI wiping out humanity, there is consensus among politicians and executives regarding the immediate threat of disinformation. Concerns focus on potential malicious use of generative AI affecting upcoming elections in various countries.
Diverse Approaches to Regulation: Different countries are moving at their own speeds in AI regulation. European diplomats highlighted their regulatory efforts, while the US emphasized its new AI safety institute. There is a shared recognition of the importance of international summits to define the problems and establish a scientific consensus about AI capabilities, similar to the International Panel on Climate Change. The UK, however, expressed hesitancy about immediate regulation due to the rapid industry advancements.