Biden said what? A summary of the 2023 Executive Order ON AI.
We get it. There is yet another important announcement on #aipolicy, but you don’t have time to keep up with it all. So, we’ve summarised President Biden’s Executive Order on Artificial Intelligence. We did our best to capture essential information, but may have forgotten a few also. If you have time, do check out the full thing. It’s a cracker.
THE summary
President Biden's Executive Order on Artificial Intelligence (AI) emphasises the U.S.'s commitment to lead in AI's promise and manage its risks. The order establishes some rigorous standards for AI #safety and #security, protect #privacy, advance equity and #civilrights, support consumers, workers, and promote #innovation and competition, whilst cementing U.S. leadership globally.
Key Directives
AI Safety & Security: Developers of influential AI systems must share safety test outcomes and “vital details” with the U.S. government. The National Institute of Standards and Technology will establish safety standards, whilst the Department of Homeland Security will oversee AI safety for critical sectors.
Privacy: There is a huge emphasise on privacy throughout. The Executive Order calls for bipartisan data privacy legislation. The focus will be on developing and using AI-based “privacy-preserving” techniques and strengthening privacy in federal agencies.
Equity & Civil Rights: Clear guidelines will be provided to prevent AI from deepening discrimination and bias. This encompasses areas like housing, justice, and healthcare. This seems like a big task, and we’ll have to wait on the guidelines to see if this has any legs.
Consumer & Worker Protection: Directives include promoting AI's responsible use in healthcare and education, whilst also addressing AI’s potential workforce impacts. So they aren’t banning AI, they are promoting a controlled use. (Implications on tax aren’t really there - I wonder if they’ve seen the idea to tax AI companies by 98% to protect workers from the possible fallout…?)
Innovation & Competition: The Order focuses on catalysing AI research nationwide, supporting smaller developers and entrepreneurs, and modernising visa processes for AI talent.
Global Leadership: Collaborations with international partners on AI safety and standards are emphasised, along with promoting responsible AI use globally. Could AI be the new forefront of transnational cooperation?
Government Use of AI: Directives are set for agencies to responsibly use AI, modernise federal AI infrastructure, streamline AI procurement, and hire more AI professionals.
This comprehensive approach, combined with international consultations, underscores the Biden-Harris administration's focus on a harmonised global strategy for AI's safe, secure, and beneficial evolution.
Considerations for Safety Tech Companies:
Transparency and Verification: With developers mandated to share safety test results, it's vital for safety tech companies to develop rigorous testing methodologies, especially for image-based AI, to detect and prevent misuse. Test. Record. Test again.
Content Authentication: The call for standards and best practices for AI-generated content identification is crucial. Safety tech companies can pioneer watermarking or authentication methods for images and videos, before it hits the public-facing platform.
Data Privacy: Image-based abuse often involves personal data. Companies should prioritise privacy-preserving techniques, ensuring images and videos aren't misused.
Collaborative Frameworks: As AI’s challenges are global, safety tech firms can collaborate internationally, sharing best practices and standards for combating image-based abuse.
Education and Awareness: As AI becomes more integrated into daily life, there’s a need for public awareness campaigns to educate about image-based AI potentials and pitfalls. Lots of safety tech companies are SaaS, but we can also play a role in this education through marketing.
Innovation & Regulation Balance: Whilst the Order promotes innovation, it's crucial for companies to ensure that rapid technological advancements do not bypass ethical considerations or regulatory standards. Soon, policy will crack down. We need to help companies be compliant, not have a non-compliance risk ourselves.
…and that’s it! What do you think about this landmark order?
This blog was originally posted on ThatsMyFace.
