資訊服務Business News

On Monday October 30th, 2023 President Joe Biden signed an executive order to adopt the safe implementation and regulation of AI. Biden stated at a press conference that “AI is all around us. Much of it is making our lives better … but in some case AI is making life worse” (Milmo & Bhuiyan, 2023). Echoing these sentiments, Vice President Kamala Harris attended the Global AI Safety Summit in the UK and advocated that governments have a “moral, ethical and societal duty to make sure AI is adopted and advanced in a way that protects the public from potential harm and ensures that everyone is able to enjoy its benefits” (Milmo & Bhuiyan, 2023). Both Biden and Harris expressed the belief that the U.S. should serve as a model for international action and leader in AI around the world.

Although Biden signed the executive order, it will still be up to Congress to ensure the safe development and use of AI. Some provisions of this executive order include requiring tech companies to share their results for their AI systems with the US government before the information is released to the public, especially if those test results could pose a threat to national or economic security, health, or public safety. Some other directives issued by the White House include 1.) setting guidelines for companies doing “red-team testing, where assessors emulate rogue actors in their test procedures” 2.) putting official watermarks on AI-made content to reduce the threat of fraud and 3.) implementing “New standards for biological synthesis screening - to identify potentially harmful gene sequences and compounds … to mitigate the threat of AI systems helping to create bioweapons” (Milmo & Bhuiyan, 2023). The mandate also covers other areas such as privacy, consumer protections, civil rights, and workers’ rights. 

Some civil liberties and digital rights groups have praised Biden for taking this important step towards prioritizing AI safeguards. Alexandra Givens, the chief of a non-profit digital rights group called the Center for Democracy and Technology, commended the administration for showing its commitment to “the responsible development and governance of AI” (Milmo & Bhuiyan, 2023). However, Givens remains cautiously optimistic because the success of the order will be based on how strictly the mandate is enforced and what actions are taken. However, others are pessimistic about the executive order, notably Albert Cahn who works in Surveillance Tech Oversight. He claims that “AI auditing techniques that ‘can be easily gamed by companies and agencies’” and even argued that most forms of AI “should not be allowed on the market” (Milmo & Bhuiyan, 2023). 

As the executive order begins to be implemented in the next few months, it hopes to also address areas such as the ethical use of AI in the military and intelligence communities. Additionally, it would call on Congress to pass legislation to protect American’s data privacy and hopefully develop guidelines to preserve privacy in AI systems. Some other areas of concern include bias in AI algorithms and trying to prevent those biases from being exacerbated. Lastly, the mandate aims to combat the threat of job displacement of workers.

Source: https://www.theguardian.com/technology/2023/oct/30/biden-orders-tech-firms-to-share-ai-safety-test-results-with-us-government