President Biden has issued an Executive Order that could redefine how the U.S. approaches artificial intelligence, outlining specific steps across multiple domains: safety, privacy, equity, and American competitiveness. It also provides a roadmap for developers, policymakers, and stakeholders, laying out actionable steps for responsible AI development and deployment.
You should read the White House Fact Sheet. Here’s a quick overview of the EO:
For safety, the order mandates sharing of safety test results for high-risk AI systems with the U.S. government. The National Institute of Standards and Technology is tasked with setting rigorous safety standards.
On privacy, the President calls for bipartisan data privacy legislation and directs federal support for privacy-preserving techniques.
Equity issues are addressed through guidelines to combat algorithmic discrimination in federal contracts and housing.
Innovation is spurred by catalyzing AI research and providing resources to small developers.
Internationally, the order advocates for global collaboration on AI safety and standards.
Federal AI infrastructure will be modernized, and a rapid hiring of AI professionals into government roles is planned.
The Executive Order serves as a comprehensive framework for AI development and deployment, focusing on actionable steps rather than broad directives. Implementation and subsequent legislation will be the metrics of its success.
From my perspective, the order is misguided political theater, adding bureaucratic overhead without efficacy. It presumes the White House can predict AI’s trajectory. Why focus solely on big tech? Small open-source initiatives could be equally (if not more) impactful or risky. The future of AI is uncertain; it will likely involve a blend of machine learning methods and yet-to-be-developed techniques.
While regulation is necessary, the order’s approach is outdated. It fails to recognize that AI innovation isn’t confined to academia or big tech. The Executive Order may stimulate industry conversation, but it’s unlikely to yield substantive regulation.
Author’s note: This is not a sponsored post. I am the author of this article and it expresses my own opinions. I am not, nor is my company, receiving compensation for it. This work was created with the assistance of various AI models, including but not limited to: GPT-4, Bard, Claude, Midjourney, Stable Diffusion, and others.