During a press conference, Alain Lebrique, director of digital and innovation at WHO, said: “ The last thing we want to see happen with this technological advancement is the spread or amplification of inequality and prejudice in the social fabric of countries around the world. ” a concern that is consistent with one of the recent IMF reports on the impact of AI on wage inequalities. With the explosive popularity of the consumer LLM, the WHO has had to update its AI guidelines in 2021, less than three years after it was first published.
WHO emphasizes that LLM “ Adopted faster than any other consumer application in history ” Capable of producing images, text or video, it is increasingly common to take advantage of it in the health sector. Our era is like the Wild West of AI; WHO fears a race to the bottom phenomenon, where a large number of companies, pressured by competition, develop poor-quality AI models. Another risk mentioned is model collapse, a great cycle of misinformation in which LLMs trained on distorted data will pollute public information sources and propagate misinformation.
Another aspect to consider is the disparity of access to the LL.M. WHO explains on page 21 of its report: “ The digital divide (…) Limits the use of digital tools to certain countries, regions or segments of the population. This divide leads to other inequalities, which affect the use of AI, and AI itself can reinforce and worsen these inequalities. ”
WHO also fears that the poorest populations will use free LLMs to access information as a substitute for real health professionals, whose services only the wealthiest individuals can afford. A big problem is that most of these models are programmed in English. ” So although they can receive input and give output in other languages, they are more likely to produce false or erroneous information. » explains the organization.
