The artificial intelligence systems are used to determine who gets public services and to decide who gets the chance to be recruited for a job, the UN chief of staff said, warning that data collected could be compromised. may be outdated and even discriminatory.
The United Nations High Commissioner for Human Rights, Michelle Bachelet, on Wednesday called for a moratorium on the sale and use of artificial intelligence (AI) systems that threaten human rights until adequate precautions are taken to ensure that the technology is not abused.
“We can not afford to continue catching up with AI — to allow it with limited or no boundaries or oversight, and to deal with the almost inevitable consequences of human rights,” Bachelet said. Press release.
The UN Office for Human Rights has a report warns Wednesday about the risks of AI technologies and stresses that while AI can serve good, it can also have catastrophic consequences if used irresponsibly.
“The complexity of the data environment, algorithms and models that underlie the development and operation of AI systems, as well as the deliberate secrecy of government and private actors, are factors that undermine the public’s significant ways of impacting of AI systems on human rights and society, ”reads the report.
Bachelet, the UN’s head of human rights, stressed that AI applications that do not comply with international human rights should be banned.
‘The power of AI to serve people is unmistakable, but AI’s ability to feed human rights violations on an enormous scale is virtually invisible. Action is now needed to protect human rights for the use of AI, for the benefit of us all, ”Bachelet emphasized.
AI’s ability to perform profiling and automate decision-making — as well as its other uses — threatens numerous human rights. It could affect the “rights to health, education, freedom of movement, freedom of peaceful assembly and association and freedom of expression”, the UN Office for Human Rights warned.
“AI systems are used to determine who gets public services, to decide who gets the chance to be recruited for a job, and of course it affects what information people see and can share online,” the high commissioner said.
The report expresses great concern about some countries and the private sector accepting AI applications without first studying the myriad risks of the technology.
There have already been some dangerous mistakes, the UN office said, where cases where people were denied social security benefits or arrested for poor facial recognition were noticed.
AI systems often collect, share, merge and analyze data in opaque ways. The information that AI collects can be compromised, outdated and even discriminatory.
“The risk of discrimination associated with AI-driven decisions — decisions that can change, define, or damage human lives — is too real,” Bachelet said. “Therefore, the impact of AI systems must be systematically assessed and monitored to identify and reduce human rights risks.”