Researchers assess global consensus on the ethics of AI applications

George Bakos/Unsplash

A team of scientists from Brazil has conducted a study to determine whether countries are ready for a unified approach to the ethical use of artificial intelligence technologies.

Here's What We Know

The researchers analysed 200 documents on AI ethics and governance from 37 countries between 2014 and 2022. These included recommendations, guidelines, policy frameworks and more.

The research team then conducted a meta-analysis of the documents. They found that the most common principles were transparency, security, fairness, privacy and accountability. The least common principles were labour rights, truthfulness, intellectual property and children's/adolescents' rights.

96% of the recommendations were "normative in nature" while only 2% recommended practical methods for implementing AI ethics. Only 4.5% suggested legally binding forms of regulation of the technology.

They also examined the geography and authorship of the papers. It turned out that most of the recommendations came from North America and Europe. South America, Africa and Oceania contribute no more than 4.5%. Male authorship is also dominant.

According to the researchers, it is necessary to attract more voices from developing countries. And also to take the next step from the principles of AI ethics to their practical implementation.

Source: Cell Press