The White House’s new executive order “Safe, secure, and reliable artificial intelligence” is poised to usher in a new era of national AI regulation focused on safety and accountability across the sector. But will it?
This executive order marks the beginning of the US government’s salvo in building a comprehensive regulatory framework for AI that applies to both the federal government and the private sector. Although it addresses broad goals for the AI ecosystem and builds on previous directives related to AI, it is not without its challenges. In particular, there is a lack of accountability and a lack of concrete timelines along with potentially overreaching reporting requirements.
This executive order perpetuates an outdated regulatory approach in which the government somehow takes the future of AI into its own hands, rather than putting in place some guardrails for the AI industry to guide itself. As recent waves of technology have taught us, development will be too fast for such an approach and will be driven by the speed of private industry. Here are my thoughts on its potential impact and effectiveness.
AI regulation
The executive order calls for the creation of new safety and security standards for AI, most notably requiring the largest model developers to share safety test results with the federal government. . But, crucially, reporting requirements remain unclear for the vast majority of companies and construction companies that are fine-tuning large-scale regulated models for specific use cases.
AI must be regulated. This is a very powerful technology, and while there is nothing inherently good or bad about it, given its power, guardrails should be put in place. While the executive order takes a focused approach to applying these standards to the largest model developers, the reporting requirements also apply to the largest underlying infrastructure providers that impact all Americans. We need to continue to mirror the progressive structure of other regulated industries as they shoulder their regulatory burdens. In contrast, U.S. regulators must have a light touch with startups to maintain the country’s leadership position in innovation.
AI security
While it’s refreshing to see the specificity of some elements, such as the Department of Commerce’s development of content authentication and watermarking guidelines to clearly label AI-generated content, many security goals remain up for interpretation. There’s room.
To quote, defense production act The order’s breadth and vagueness of language set it on the path to overreach, as it prioritizes “sharing test results and other sensitive information with the U.S. government” that could nationalize companies in an emergency. may open. While it may be touted as championing American values, the potential for overreach is clear. This section of the executive order includes whimsical statements such as, “The Department of Energy and the Department of Homeland Security will address chemical, biological, radiological, nuclear, and cybersecurity risks, as well as threats from AI systems to critical infrastructure.” A description is also included. Vague language and scope of the mandate leave room for ambiguity, which could lead to unintended consequences.
Protecting consumer privacy
The AI Executive Order also calls for protecting consumer privacy by creating guidelines for government agencies to evaluate privacy technologies used in AI. If authorities adopt these guidelines, they could be helpful for entrepreneurs who want to avoid conflicts with regulators. Of course, such guidelines are only effective in fostering innovation when applied consistently.
Similarly, the executive order calls on the Department of Health and Human Services to develop programs and resources on how educators can use AI tools responsibly. This use case-specific regulatory approach is effective because it allows regulators to apply existing frameworks to guide innovation in these areas.
open code
The executive order’s disturbing details require AI companies to “pull back the curtains” and share their products’ internal test data with the National Institute of Standards and Technology. The group is designated to set safety standards for “red team” testing to ensure safety before public release.
Given the speed of innovation, it makes little sense to regulate AI research and development. Instead, it’s best to apply a light touch to research and development and adapt as innovations unfold. Therefore, further clarification is needed regarding detailed R&D reporting.
Recruiting AI talent
Perhaps the most encouraging aspect of the executive order is that the existing They are calling for liberalization of visa standards, interviews, and examinations. America.
This includes relaxing strict H-1B work permit standards for non-U.S.-born workers. This executive order aims to ease the immigration of highly skilled workers, which is a key element in the continued growth of America’s AI companies, research, and professional workforce.
5 years have passed since I I addressed this issue And how to attract and retain fresh talent, educators and data scientists must be part of the national agenda. I believe those words even more strongly today.
Immigration has been a driving force behind America’s technology startups. A 2022 study found that 55% of U.S. startups (valued over $1 billion) were founded by immigrants. National Foundation for American Policy Analysis. More importantly, 64% of billion dollar unicorn companies were founded/co-founded by immigrants or their descendants.
The same think tank recently estimated that 65% (28 of 43) of the top AI companies in the US were founded or co-founded by immigrants, and 42% of founders came to the US as international students.
These numbers clearly demonstrate that international entrepreneurs are contributing meaningfully to the technology/AI industry, and it is in the national interest to continue to attract and retain new talent for leadership positions in AI. I am.
It is important to invest in existing students along with foreign workers. This executive order strengthens his AI research by providing students and AI researchers with access to essential AI resources and data, and provides technical assistance to small and medium-sized enterprises, the National AI Research Resource was established.
These students will become the next machine learning experts, AI algorithm engineers, and tomorrow’s company founders who can help unlock the full potential of AI and advance America’s innovation and leadership. These are encouraging steps to accelerate the development of this fast-growing industry.
This executive order represents a vital step in regulatory and AI advancement in the United States. However, this has challenges and ambiguities and requires further scrutiny and refinement.