Trump Should Keep His New AI and Crypto Czar, David Sacks, Accountable for Responsible AI Development
December 6, 2024
While President-elect Donald Trump is widely expected to follow through on his promise to “repeal” President Biden’s executive order on AI, the order contains a number of provisions that are consistent with Trump’s past statements on responsible artificial intelligence development.
Voters, meanwhile, overwhelmingly support government guardrails on AI development. According to a recent poll from the AI Policy Institute, “80% of voters—including 76% of Republicans and 84% of Democrats—prefer a regulatory approach that mandates safety measures of the most advanced AI models and government oversight for the certification and release of new AI models.”
President-elect Trump and his new AI and Crypto Czar, David Sacks, should consider pushing for the codification of the provisions in President Biden’s executive order that are consistent with President-elect Trump’s stated goals.
Trump’s primary justification for pulling back on AI regulation stems from a desire to keep the U.S. at the forefront of AI innovation and particularly, to counter China’s influence. Nevertheless, in past statements, he and some of his allies have underscored the need to govern the safe deployment of AI while encouraging its rapid evolution.
The following are four tenets of Biden’s AI order that are aligned with Trump’s interests, based on the latter’s past comments:
- Developing technology to label deepfakes and other potentially harmful synthetic content: Manipulated content impersonating another person or entity can erode copyright and privacy protections, increasing the potential for fraud. Both Biden and Trump have called for labelling of synthetic content and the development of advanced techniques, such as watermarking, to enable viewers to identify whether content is original. Specifically, Trump has expressed concern about potentially harmful synthetic content, describing with unease in an interview with Logan Paul a potential scenario where a foreign adversary is sent an AI-manipulated video suggesting that Trump has launched a nuclear missile, influencing that leader to launch a preemptive strike.
- Calling for transparency and coordinating data sharing: One of Trump’s most pointed criticisms of Biden’s executive order centers on the requirement for companies that develop large-language models (LLMs) to share detailed information and data on their development activities with the federal government. But both presidents have called for transparency and data-sharing more broadly. American conservatives have expressed concern over the last few years that AI models might be biased against them. President Trump himself has shared similar concerns regarding social media companies’ content moderation systems. Ensuring that AI models remain fair for all people, including conservatives, would involve some level of oversight on model training and testing.
- Advancing export controls to prevent foreign manipulation: Both presidents have said that the U.S. must do what it can to protect U.S. AI intellectual property and prevent foreign adversaries from manipulating AI to the detriment of society. Both presidents have also placed export controls on China for key AI inputs. During his first term, Trump issued two executive orders containing ample material with respect to national security as it relates to AI.
- Enabling an AI-ready workforce: Both presidents have said that the workforce will need to be trained and re-trained to capitalize on AI and to prevent high rates of structural unemployment.
A pullback in AI regulation does not have to mean total inaction with respect to AI safety. President-elect Trump needs to consider his priorities with respect to responsible AI development and assign Sacks to advocate for the codification of those priorities into laws and executive rules while simultaneously investing in continued innovation.