Artificial Intelligence and Human Rights — Why does it matter?

By Harley-Jane Dority, Government Relations Intern, Amnesty International USA

We live in a rapidly-evolving world where new developments in science and technology are constantly claiming to make our lives easier and more productive. Artificial intelligence (AI) is one such leading technological trend, with AI systems seemingly promising efficiency, neutrality, and in many cases the solutions to long-standing social and economic problems — in sectors as diverse as healthcare, finance, law enforcement, and employment.

Amnesty International defines AI as advanced computer software and computer-powered hardware that can undertake self-learning computational or physical tasks. It sounds futuristic, but AI technology is in fact everywhere and you’re likely using it every day if you ever take an Uber, utilize Google Maps, deposit a mobile check, or scroll through social media sites like Facebook and Snapchat. When a ride-share application determines the price of your ride or Facebook tags a friend in your recently-uploaded photo before you do, AI technology is behind those actions.

One may ask “What does this have to do with my life?” Self-driving cars are a cool concept but you’re still commuting to work in a 2012 Honda Accord, so why should you care about the fast-paced technology associated with AI and the human rights concerns that accompany it?

While these innovations are making our lives easier in some ways, it is crucial that AI technology and current systems consider the impact of AI on human rights, and that governments and companies respect current human rights laws and standards even in the use of new technology.

The Lantos Commission for Human Rights held a hearingon May 22 focusing on AI and the consequences for human rights. Co-Chair of the Commission, Rep. Randy Hultgren, thanked Amnesty USA at the hearing for the statementwe released for the record. Amnesty USA offered a summary of recommendations to the US government and US-based companies that largely drive the global development of AI technologies, asking them to consider human rights and act accordingly as development continues in the field. We need to make sure that human rights and the long-established mechanisms for holding rights abusers to account are seen as relevant and applicable for abuses in this technology — and, crucially, that they are used to protect rights at risk of harm.

Rep. Hultgren retweeted a quote from the Lantos Commission in which Mark Lantonero statedthat,“The application of big data in the human rights domain is still really in its infancy. The positives and negatives are not always clear and often exist in tension with one another, particularly when involving vulnerable populations.” Sherif Elsayed-Ali, Amnesty’s Director of Global Issues, spoke tothe need for Amnesty to have a proactive and deliberate approach to the impact that new technologies were having on human rights. Amnesty recently launched The Toronto Declarationprecisely for this reason. An open Declaration from Amnesty and digital rights group Access Now that has been widely endorsed by human rights groups, The Toronto Declaration outlines the roles of states and companies in ensuring that the right to equality is not undermined by AI technology.

So, how does this affect you? You use social media, input personal information into apps, and utilize email communication systems. As AI advancements continue, vast economies of personal data are developing, creating networked systems that store and process that data to an unprecedented degree. Why does this matter? AI systems may lead to privacy and profiling risks in the collection of personal data. Networked big data may be used to create intimate and precise personal profiles of individuals, a tactic already widely used for commercial advertising and political marketing during elections. AI software makes profiling on such an intimate, individual level much more accessible — with the potential for companies and governments to influence people to a greater degree than ever before. Limiting freedom of choice is a problem, but even more so, the discriminatory power of algorithms and the threat to equality both through intentional discrimination and inadvertent discrimination is a major concern in the development of AI technology.

While AI technologies have the potential to improve access to human rights, for example, helping to widen access to advanced healthcare diagnostics and treatments, the need for adequate regulation, transparency, and accountability is critical if the field of AI is to take human rights, equity, and protection seriously.