A House Democrat is warning artificial intelligence could become a tool of “digital colonialism” if the U.S. doesn’t take steps to work with Western Hemisphere nations to create AI systems that reflect diversity and inclusion.
Rep. Adriano Espaillat, D-N.Y., proposed a resolution during the August break that says the U.S. must champion a “regional” AI strategy that includes Western Hemisphere nations as this new technology is developed.
“United States-led investments in the development of AI in the Western Hemisphere would promote the inclusion and representation of underserved populations in the global development and deployment of AI technologies, ensuring that no individual country dominates AI but rather collaborative developments in the Western Hemisphere,” his resolution asserted.
Without naming China, it hints that allowing authoritarian regimes to take the lead on AI standards would only hurt vulnerable populations in the Western Hemisphere.
“The United States’ future policies for AI governance will have significant implications for the global governance of AI, influencing whether global AI technologies reflect democratic values, inclusivity, and respect for human rights or are influenced by authoritarian practices and norms, including ‘digital colonialism,’ whereby most of the AI advancements utilized by the Western Hemisphere consumers would be developed in, and controlled by, a select few nations located outside of the region,” it warned.
Espaillat, born in the Dominican Republic, said American efforts to work with the rest of the Western Hemisphere on AI standards would “contribute to a more equitable, responsible, and human-centric approach, ensuring the development and deployment of AI technologies that align with democratic principles and societal well-being.”
“By championing inclusion in AI and investing in AI in the Western Hemisphere, the United States can create a future where AI technologies authentically reflect the multifaceted diversity of our societies, uphold the fundamental human rights that lie at the core of our Constitution, and contribute to the realization of a world that transcends inequalities rather than perpetuates them,” the resolution said.
“The Western Hemisphere possesses a wealth of natural resources and a skilled human workforce, making it well-positioned to develop and promote future AI technologies that prioritize safety, diversity, equity, inclusion, and accessibility,” it added.
It warned more broadly that AI has the potential to be developed in a way that reinforces “biases and inequities.” Others have made the same argument – that AI systems run on data that is either biased or interpreted in a biased way can lead to outcomes that discriminate against what Espaillat and others call “marginalized groups.”
Espaillat’s resolution said recent research shows some AI algorithms can worsen “race-based disparities,” such as those used in facial recognition programs.
“Research conducted by respected institutions such as the Institute of Electrical and Electronics Engineers, the Massachusetts Institute of Technology, Cornell University, and others has shown that these algorithms exhibit significant accuracy disparities, working more effectively on White faces while frequently misidentifying or failing to recognize brown, Black, Indigenous, and darker skinned faces,” it warned.
It’s not clear Espaillat’s resolution will get a vote in the GOP-led House where it was introduced. However, the Biden administration has launched several initiatives aimed at developing safe and trustworthy AI systems that avoid biased outcomes as they are used.
Last month, seven major AI developers agreed to a set of White House goals in this area, and the Biden White House has said it is working on more AI guidance.