The revolutionary power of computers to “learn” from large datasets and use this information to make predictions is unprecedented. Loosely defined under the umbrella of artificial intelligence (AI), these complex systems are already applied in multiple contexts: at home, at work, in decisions about loans and college admissions, in vehicles and transportation systems, in weapons, in surveillance, in healthcare, and in policing and judicial sentencing. Future applications of AI could extend even further, encompassing most aspects of our lives.
As we stand on the cusp of a technological revolution, opportunities abound to harness these technologies for social good: AI-based systems may have the capacity to scan social media across platforms and across the globe to warn of imminent environmental disasters, disease outbreaks, or political upheavals, using this information in ways that individuals may not have intended or even realized were possible. They have the potential to remove human bias from judicial decisions about sentencing; AI-based systems could optimize diagnostics and therapeutics through analysis and correlations of immense swaths of patient data.
But for the benefits to be equitably enjoyed by every member of society, public attitudes toward AI among marginalized and vulnerable populations must be better understood. AI innovations must be intentionally and responsibly developed, and their application must address the possible disparity of impact new technologies have on disadvantaged populations. In response to these challenges and opportunities, the American Association for the Advancement of Science (AAAS) and its partners are launching the (AI)²: Artificial Intelligence—Applications/Implications initiative, the goal of which is to affect the development and application of AI in ways that alleviate, rather than exacerbate, social inequalities.