The White House has declared a set of measures that it deems as “the most substantial actions ever undertaken by any government to promote advancements in AI safety.” President Joe Biden’s executive order mandates that developers of Artificial Intelligence (AI) must disclose safety outcomes to the US government, positioning the United States at the forefront of the worldwide discussion on AI governance. Concurrently, the UK government, under the leadership of Prime Minister Rishi Sunak, is conducting a summit on AI safety this week.
Commencing on November 1 at Bletchley Park, the two-day gathering has been prompted by apprehensions about the swift progress of AI systems, which could potentially result in issues like the creation of more lethal bio-weapons and more debilitating cyber-attacks. In unveiling these safety measures, Mr. Biden pledged to “utilize the capabilities of AI while ensuring the safety of Americans.”
Insufficient Teeth
According to tech entrepreneur and AI expert Gary Marcus, the US announcement appears more ambitious in its breadth. Marcus stated that Biden’s executive order establishes a significant initial standard, encompassing both immediate and future risks, though he remarked that it may lack the necessary robustness.
“The UK summit seems to have considerably narrowed its focus, primarily emphasizing long-term risks, with insufficient attention on the current challenges. It remains unclear how much enforceability or authority will be derived from it,” Marcus commented.
Alex Krasodomski, a senior research associate at Chatham House, noted to the BBC that the executive order signals the US sees itself as a leader in addressing such threats.
During a White House address on Monday, President Biden stated, “As artificial intelligence expands the boundary of human possibility and tests the bounds of human understanding, this groundbreaking executive order is a testament to what we stand for: safety, security, trust, openness, American leadership, and the undeniable rights endowed by our creator that no creation can take away.”
The US measures include:
- Establishing new safety and security standards for AI, with requirements for AI companies to share safety test results with the federal government.
- Safeguarding consumer privacy by developing guidelines for agencies to assess privacy techniques used in AI.
- Combating AI algorithms that lead to discrimination and formulating best practices for the proper role of AI in the justice system.
- Creating a program to assess potentially harmful AI-related healthcare practices and offering resources on responsible AI tool use for educators.
- Collaborating with international partners to implement AI standards globally.
Additionally, the Biden administration is enhancing its AI workforce by opening federal government positions related to AI expertise on AI.gov starting Monday.
Does this Counteract the UK Assembly?
Mr. Krasodomski acknowledged the significance of the executive order but pointed out a misalignment with the UK’s objectives for the summit. He emphasized that the UK summit is referenced in the order within the context of “advancing American leadership abroad,” signaling the US’s awareness of its major role, particularly in comparison to China. He noted that it is the US companies that are driving advancements in this field.
“It’s challenging to organize a small, highly technical summit, but it’s evident that for this technology to have a substantial global impact, extensive collaborative efforts and engagement with countries worldwide are imperative,” Mr. Krasodomski remarked.
Despite this, US Vice-President Kamala Harris and top executives from major US tech companies are scheduled to attend the UK government’s AI Summit this week, touted as a “world first” in addressing AI safety.
The summit is set to address escalating concerns surrounding the ramifications of what is termed as frontier AI. Noteworthy figures such as President of the EU Commission Ursula von der Leyen and UN Secretary-General Antonio Guterres are slated to participate.
The UK is resolute in establishing itself as a global leader in mitigating the risks associated with this influential technology. However, the EU is currently progressing with an AI act, China has already implemented stringent AI regulations, and the US has recently issued a significant executive order.
Moreover, Reuters reports that the Group of Seven (G7) industrial nations is purportedly in the process of formulating a code of conduct for companies engaged in the development of advanced AI systems. With this heightened global activity, there arises a question about the extent to which substantive discussions will take place at Bletchley Park this week.
