Google added that part of the problem it faces when generating AI summaries is the lack of quality information on the web for very specific queries. And there is little doubt that Lin’s work is not of high quality.
“The quality of the science underlying Mr. Lin’s ‘National IQ’ database is so poor that it is difficult to believe that the database is anything other than a scam,” Shear said. “Lin never explains the methodology for selecting samples for the database. Many countries estimate IQ from absurdly small and unrepresentative samples.”
Shear pointed out that Lin’s estimate of Angola’s IQ was based on information from just 19 people, and that Eritrea’s IQ was based on a sample of children living in orphanages.
“The problem is that the data that Lin used to generate this dataset is just bullshit, and it’s bullshit in many ways,” Rutherford said, adding that the numbers for Somalis in Lin’s dataset are It noted that it was based on a single sample of refugees aged 8 to 15. 18 people tested in Kenyan refugee camp. He added that Botswana’s score was based on a single sample of 104 Tswana-speaking high school students aged 7 to 20 who took the test in English.
Critics of using national IQ tests to promote ideas of racial superiority argue that not only are the samples collected of poor quality, but that the tests themselves are typically designed for a Western audience. It is also pointed out that there is a bias in advance because of the Even if administered.
“There is evidence that Mr. Lin systematically biased the database for African countries by preferentially including high-IQ samples, while preferentially including high-IQ samples,” Sears added. , a conclusion supported by a 2020 preprint study.
Lin has published various versions of the national IQ dataset over the decades, the most recent of which was published in 2019, called “The Intelligence of Nations.” Lin’s flawed work has been used by the far right and racists for years. Groups as evidence supporting claims of white supremacy. The data is also presented in a color-coded world map, with countries in sub-Saharan Africa, which are thought to have low IQ, shown in red, while Western countries are shown in blue.
“This is data visualization that you see all over Twitter and social media, and for anyone who spends a lot of time in racist hangouts on the web, this is just a way of saying, ‘Look at the data.’ It just looks like a racist argument. Look at the map,” Rutherford says.
But Rutherford believes the blame lies not just with AI systems, but with the scientific community that has cited Lin’s work uncritically for years.
“It’s actually not surprising[that AI systems would cite it]because Lin’s work on IQ has been accepted without question from vast swaths of the academic world, and his national If you look at the number of times the IQ database has been cited in academic papers, it’s in the hundreds,” Rutherford said. “So the responsibility lies not with AI, but with academia.”
