US announces most powerful AI safety act of any government to date

2 Nov 2023

Written by Emma Hall (Digital Editor)

With the US recently releasing the most powerful act we have seen yet for AI safety, can we finally spot global AI regulation on the horizon?

The US has released an executive order for AI safety this week, signed by Joe Biden at the White House on Monday. The new directive is said to be the “most significant actions ever taken by any government to advance the field of AI safety” and aims to take optimum advantage of AI potential while protecting US citizens from harm.

The executive order addresses both the long-term and short-term risks that AI poses, and includes several key policies and principles:

Creating appropriate guidelines for AI developers to create safe, secure and trustworthy tools. This involves requiring AI developers to share safety test results with the US government.
Collaborating with international allies and partners to develop a framework to mitigate AI’s risks with common solutions, encouraging a unified global approach in creating responsible AI.
Shielding consumer privacy by ensuring appropriate data collection, including using privacy-enhancing technologies where possible.
Addressing AI biases and discrimination by inspecting algorithmic performance, identifying discrimination in current systems and assessing appraisal processes.

The directive also includes information regarding AI safety in healthcare. This consists of creating a Health and Human Services (HHS) AI Task Force, supporting initiatives that improve health data quality, assessing the pre-market and post-market performance of AI healthcare tools against real-world data, and developing a standardized method of pinpointing clinical errors resulting from AI use.

Closely following suit, the UK is set to host the world’s first global AI Safety Summit this week from 1–2 November (Bletchley, UK). Arranged to discuss and mitigate the potential risks that AI poses, the summit aims to stimulate international conversation over global regulation, something that has been particularly slow in response to the pace of AI advancements.

The group of global policy leaders and AI experts will gather to consider matters such as AI weaponization, cybersecurity, and human rights infringement. British Prime Minister Rishi Sunak will be summarizing the meeting discussions later today in a speech when the summit ends.

Last week, the UN also declared that it has established its own advisory board, comprised of experts from research, industry and governments.

With global efforts towards AI safety finally beginning to take off, we hope to soon hear of more unified regulation. For now, we wait with bated breath.