Human rights in an era of artificial intelligence

Hardly a day goes by without another story on Artificial Intelligence, or AI. Heralded as a technological breakthrough with enormous potential to transform human life, fear about its risks is on the rise. Indeed, it has become yet another battleground between the United States and China.

Technological change has both sustained and harmed humanity from its beginning, even to deadly effect. Modern technology is no different. Many inventions have enriched our lives. Some have also promoted evil ends. What is different today is the speed of technological change.

Human rights are at particular risk, under siege around the world. One reason for the recent deterioration of those rights, tragically, has been the rise of the internet and associated technologies. Of course, the online revolution was supposed to be a force of global liberation, and in some ways it is. However, it also generated China’s Great Firewall. Authoritarian forces deploy software bots to disrupt and pervert political debate in democratic states. Beijing has ruthlessly coopted internet providers.

And now we are being buffeted by the rise of AI.

No doubt, it will do much good. However, there also are many ways in which AI will be used to violate human rights and undermine democracy. We are, unfortunately, only at the beginning of what will almost certainly be a long and difficult battle. Autocracies will not be alone when it comes to the misuse of AI. Activists and government actors in democracies will also be tempted, especially as politics grows more bitter, partisan, and confrontational.

For instance, even with the best of intentions, use of AI might introduce subtle bias in speech moderation by internet platforms and others. Most systems are likely to have problems with the context of argument and nuance of language. Worse, algorithms could be designed to manipulate, with distortions disguised and difficult to discern. The intent could be to suppress, intimidate, or amplify particular viewpoints and/or participants. If people grow to distrust mainstream platforms, they are likely to frequent more extreme sites and publications.

AI could be used to intensify illicit and inappropriate discrimination, disproportionately directed against already vulnerable and targeted groups. The lack of transparency regarding systems’ operations would help shield them from scrutiny. Imagine technology that purported to assess future beliefs and behavior, leading to “predictive policing.” Government or other malign forces will begin to take preemptive action to eliminate perceived threats. Often victims may not even have been aware of how they were targeted.

Governments and private actors could use AI to entrap “undesirables.” AI would make it easier to bait a trap for political dissidents and others, setting them up for extortion, punishment, or worse. Threats against freedom to associate and assemble would grow more intense.

Of particular concern to rights advocates is the potential for AI to intensify and expand surveillance, especially through facial recognition. AI could be used against both individuals and groups. At a minimum, when employed for surveillance, it is likely to have a massive chilling effect. In this way, AI would depress both legitimate political protest and normal social life. It also could target communities or entire regions as part of a campaign of political, racial, or ethnic persecution.

AI is likely to make it easier to craft misinformation to mislead and discourage voters. The aim could variously be to change votes, degrade civic debate, create social mistrust and/or discourage political participation. Bots and synthetic media, especially deep fakes, already undermine American democracy. AI is certain to exacerbate this threat.

The rapid rise of an online society has harmed people’s privacy. AI is likely to accelerate this process, especially through constant tracking, which would both strengthen existing tools and likely provide new means for gathering, disseminating, and employing data. In this way AI could empower governments and even private actors, both domestic and foreign, to exploit people’s personal problems and vulnerabilities.

Malign actors also could use AI to expand and intensify censorship. They could assess who might threaten the regime and what issues might inflame opposition. This process would discourage legitimate opposition while tagging otherwise innocent people as likely dissidents and innocent topics as seditious commentary.

Governments could use AI to tie together many of the forgoing actions into an all-encompassing “social credit” system that regulated even the most personal decisions and delivered benefits and exacted punishments in response. The danger posed by authoritarian regimes such as the CCP is obvious. But even intrusive western governments could also find such a system to be attractive.

More broadly, turning decisions over to machines would dehumanize public and civic life. The belief, even if more exaggerated than real, that private as well as public discourse is directed by others, especially by computers, programs, and algorithms, could create widespread cynicism, discourage public participation, and even weaken nonpolitical social life. Ultimately, this process could undermine individual responsibility and accountability, with a destructive impact on people’s relations with one another.

We are truly living in perilous times. There is danger, to be sure. However, our greatest problems lurk at home, sometimes in unlikely and confusing places, such as AI. Unfortunately, there is no simple solution.

As a first step, we must recognize the dangers, which are real and serious. Concern for human rights must not be lost in a modern “gold rush” fueled by the irresponsible greed of investors or the thirst for social control by the state and other powerful institutional actors. Defending human rights will require an international effort, but we must begin at home.

The next step will require that we work past our disagreements for the common good. Those who value freedom can and should come together around the need to protect human rights from a litany of new dangers in the era of AI. Who else will defend our free society or the democratic principles upon which it rests?