With its ability to simplify our everyday lives, artificial intelligence (AI) has earned itself a central role in modern society. Yet, in the absence of adequate regulation, the clash between AI and human rights is increasingly evident. From fostering discrimination to enabling invasive surveillance practices, AI constitutes a threat to basic freedoms, equal protection and socio-economic rights.
What is AI?
“AI” refers to algorithms that simulate human intelligence, by mimicking behaviours like ‘learning’ and ‘problem solving’. What differentiates AI from the computer programmes we used a decade ago, is that AI changes itself, without needing human input. While ‘normal’ computer programmes were written once - and would carry out the same tasks over and over again - AI-based programmes will re-write themselves over time. For example, after a while the autocorrect on your phone will register that you never want to change ‘pingpong’ to ‘pinpoint’, and will stop offering that possibility. While not being harassed by your phone to change pingpong to pinpoint is certainly a good thing, this is not true of all the things computers learn.
How does artificial intelligence affect our lives?
AI as a tool of discrimination
AI-based facial recognition systems have repeatedly shown to be less capable of identifying Black people than Caucasians. Instances of “racist” AI are ubiquitous online. In 2015, Google Photos, which was widely considered an advanced facial recognition software, notoriously labelled two Black individuals as gorillas. Equally degradingly, the AI-determined Google search results for “black girls” are almost entirely sexually explicit. Such biases may lead to false arrest, missed flights, and all sorts of other difficulties for Black people and help perpetuate negative stereotypes and stigma.
The use of AI in law enforcement in a growing number of countries, including, Hong Kong, China, Denmark, US and India, also raises a host of discrimination concerns. AI is often fed data produced by biased policies like over-policing of certain ethnic minorities. This data would indicate that crime rates tend to be higher among these groups, who are then more likely to be classed as high-risk offenders by algorithmic systems. This is blatantly unjust. It decreases the chances of illegal behaviour by someone from the majority population being detected, while entrenching the overrepresentation of minorities in the criminal justice system. This then creates broader social and economic inequalities for people from minority communities.
AI at war
Developments in AI have also changed the military sector. Armed drones are increasingly used in war, and some of them are not controlled by humans at all anymore. While “killer robots” not so long ago seemed to be in the semi-distant future, technology developed faster than expected. The inability of AI to make ethical judgements raises questions about whether the use of such weapons might result in accidental deaths and the rapid escalation of conflict.
Technology as a source of unemployment
We are accustomed to seeing new technologies change existing employment and power structures. But AI will bring a particularly big shift.
Already, AI has set off a trend of increasing rates of unemployment. For instance, Changyin Precision Technologies – a Chinese factory that produces mobile phones – has replaced 90% of its human workforce with machines, leading to a 250% increase in its productivity and an 8% drop in defects. Adidas has similarly moved towards ‘robot-only’ factories to improve efficiency (though, it seems, with significantly less success so far). Gloomy prognoses about the automation of work indicate that eventually a human workforce might become redundant or, worse, detrimental to business growth. Robert Skidelsky, an Oxford historian, writes that “sooner or later, we will run out of jobs” as 9%-47% of jobs are estimated to be at risk of automation in the near future. Indeed, even jobs that we might consider immune from automation can be replaced by AI. Just think of AI-based virtual assistant software, such as Siri, Cortana, Alexa and Google, which have steadily replaced personal assistants, translators, and other services previously provided by people. At present, AI mostly threatens the jobs of people who are unskilled and low-skilled, who will find it increasingly difficult to make a living.
Governments are under an obligation to ensure an adequate standard of living for their populations. This usually translates into stimulating fair-paying jobs in combination with social services and support for people who are unable to work. If employers are allowed to substitute their workforce for AI-controlled machines, governments will have to come up with alternative ways of providing for their peoples. For example, some have suggested an AI tax to pay for a universal basic income. Making sure people have an income is not just to allow people to live in dignity. History has shown us that unscrupulous politicians can capitalise on mass unemployment to radicalise voters, create authoritarian regimes, and persecute scapegoated groups. .
To some, AI carries the promise of incredible efficiency. But it poses great challenges to our freedoms and standard of living. Lawmakers will need to regulate AI carefully so it becomes a tool to enhance, rather than threaten, our values and quality of life.