In a wide-ranging Executive Order issued this week, President Joe Biden sought to preempt threats posed by the development of artificial intelligence with new safeguards and oversight, while calling on Congress to pass data privacy legislation and enshrine such guardrails permanently.
While AI has become a hot-button word, guiding ongoing developments at major companies like Google and Amazon, and embodied in controversial bots like ChatGPT, the White House has also warily eyed its potential for developing devastating weapons and launching new cyberattacks. By tapping executive authority under the Defense Production Act, Biden crucially ruled that AI companies must conduct safety tests of some of their products and share those results with the federal government before opening them up to consumers.
While specific enforcement was lacking in the order, it did promote a sense of urgency on the topic, focusing on the present danger of AI. Unlike in movies, it’s not AI becoming too self-intelligent that has sparked fears, but how people are and will use it.
“One thing is clear: to realize the promise of AI and avoid the risks, we need to govern this technology, and there’s no other way around it, in my view,” Biden said in remarks at the signing. “It must be governed.”
The Department of Homeland Security (DHS) will be tasked with creating an AI Safety and Security Advisory Board (AISSB) to support responsible development of AI, composed of industry experts. The White House wants its recommendations and best practices to support AI development in secure ways. At the same time, DHS will work with stakeholders to create safety and security guidance for use by critical infrastructure owners and operators, building on work by the Cybersecurity and Infrastructure Security Agency (CISA) to assess potential AI-related risks to critical infrastructure sectors.
While keeping it from nefarious uses, the administration also wants to tap into AI’s potential power for improvements, specifically to U.S. cyber threat detection, prevention and vulnerability assessment capabilities.
“The unprecedented speed of AI’s development and adoption presents significant risks we must quickly mitigate, along with opportunities to advance and improve our work on behalf of the American people,” Homeland Security Secretary Alejandro Mayorkas said. “The President’s Executive Order defines a seminal path for the safe and secure use of AI. It directs DHS to manage AI in critical infrastructure and cyberspace, promote the adoption of AI safety standards globally, reduce the risk of AI’s use to create weapons of mass destruction, combat AI-related intellectual property theft, and ensure our immigration system attracts talent to develop responsible AI in the United States. The AI Safety and Security Advisory Board, which I look forward to chairing, will bring together industry experts, leading academics, and government leaders to help guide the responsible development and safe deployment of AI.”
Concerns also extended to intellectual property theft – the focus of several ongoing, high-profile lawsuits against AI companies. The order reiterated that IP theft threatens U.S. businesses and security, and announced a new program to help AI developers reduce AI-related IP risks – something that has frequently emerged as AI reaches out and processes data from abroad, with little concern for sourcing. Watermark efforts will focus on reducing the threat from deep fakes and stolen material, and to up transparency.
The EO also made it clear the concern with which the federal government looks at AI’s potential in the area of weapons of mass destruction (WMDs). This is particularly true of AI-enabled misuse of synthetic nucleic acids, which could be used to create biological weapons. DHS will guide work to create a framework of evaluation and stress testing of these sorts of screens, and to develop a standardized set of expectations for third party auditors of AI systems.
Still, much of these measures will act as suggestions rather than law, which is why the White House called for action from Congress as well. Cultivating talent was also highlighted as a major focus going forward, including reworking of certain immigration pathways to get the right talent in and working on U.S. AI technologies.