The administration supported a race for AI when it could have supported a race for ethical innovation.
The White House and the Pentagon can feel the hot breath of artificial intelligence competition from China on their necks. In its desperation to ensure that the U.S. remains a dominant force in AI development, the White House missed an opportunity in its executive order on AI strategy to finally go on the record and address the ethical questions of artificial intelligence in a meaningful way.
The Pentagon, home to development of AI weaponry and other tools that are so much cause for consternation among ethics advocates, mentioned ethics in its own AI strategy but buried it beneath promises of rapid adoption and experimentation.
The White House strategy bucks the AI ethics trend happening across the globe as countries such as Finland, Dubai and Singapore establish their own more-thoughtful approaches to AI development. Policy-makers have formed committees and invested time and resources into devising national guidance for ethical AI. Meanwhile, the U.S. has stood on the sidelines of the AI ethics movement, failing to address the real-world impact AI technologies have on people and society.
There’s a reason the AI ethics movement is resonating: repeated headlines about predictive policing tech that perpetuates discrimination, poorly-crafted automated credit scoring tools, and dwindling work opportunities.
But the Trump administration is focused on defeating China, our closest AI industry competition. The executive order itself directs the assistant to the president for national security affairs to “organize the development of an action plan to protect the United States advantage in AI and AI technology critical to United States economic and national security interests against strategic competitors and adversarial nations.”
The Defense Department’s AI strategy reinforces concern that military investments in AI in China and Russia “threaten to erode our technological and operational advantages.” And it merely tips a hat to ethical AI development and financial support for work developing more explainable AI systems. Instead, the department’s “central focus” of its AI strategy, reiterated in its own press release is “increasing speed and agility.”
Countries Across the Globe Couple AI Innovation and Ethical Goals
Countries around the world have tempered AI ambitions with recognition that taking a move-fast-break-things approach to AI could have irreparably detrimental impacts on society and humanity as we know it. Since at least 2016, national government AI strategy communications have addressed ethical considerations, many in great detail.
Over the past several months, I have analyzed 22 national government AI strategy directives and reports from the EU, U.K., Asia and the Middle East, supplemented by research on European government efforts from AccessNow.
These countries all made efforts to discuss the importance of establishing ethical principles. Most mention the need for transparency in automated systems, enabling explanations for decisions that directly affect people. Many emphasize the importance of AI fairness to avoid systems that discriminate against minorities or protected groups. Several address the need for privacy protections and data use limits in data-hungry AI development.
Japan’s strategy stressed issues including accountability and liability, along with a mission to ensure diversity among those creating AI systems. Singapore’s AI Governance Framework published in January called for transparency and emphasized the importance of human autonomy. France, Germany and Finland addressed the potential effects of AI on the labor force and workers’ rights.
Meanwhile, in discussing “AI and the American Workforce,” Trump’s executive order fails to mention the ethical concerns swirling around the future or dignity of work in an AI-enhanced world. Instead, it limits the discussion to workforce training.
The order refers to the need to protect “civil liberties, privacy, and American values,” but with no elaboration or hint of irony regarding our country’s failure to pass federal privacy legislation over the past decade.
While other countries have led the way by crafting substantial reports and declarations regarding strategies for AI that detail ethical considerations, Trump’s executive order took a hasty and directionless approach. The administration supported a race for AI when it could have supported a race for ethical tech innovation.
Kate Kaye is tech journalist focused on AI ethics for RedTailMedia.org.