Researchers at the AI Now Institute have raised an important question about the direction of this area in today's report that the artificial intelligence industry is facing a "diversity crisis."
Women and colors were found to be scarce. According to the report, about 80% of AI professions are male, but 15% of Facebook AI researchers and 10% of Google were women. People in color are out of the picture, with only a handful of key IT staff. The result is often a man-driven workforce driven by white and male perspectives that affect people in other groups. "This is not the diversity of people affected by this system," says Meredith Whittaker, co-director of the AI Now Institute.
The plan to amend the potential pipeline "pipeline" Despite decades of "pipeline research" to assess the diverse flows of job candidates from school to industry, there has been no significant progress in the diversity of the AI industry.
The researcher makes suggestions for improving the problem.
Diversity is a hurdle in the high-tech industry, but it presents specific risk factors for AI, but companies say they can improve transparency by posting more compensation, racial and gender-specific data, and posting bullying and discriminatory transparency reports . Where potentially biased technologies such as face recognition can disproportionately affect historically marginalized groups. Tools such as a face scan program to determine the sexual orientation introduced in 2017 reflect the absurdities of the past. Rigorous testing is required. But more importantly, those who make artificial intelligence tools are not willing to create dangerous projects. Kate Crawford, co-director of the AI Now Institute, said: "We need to know that these systems are not only fair, safe, and secure.
Technology workers can address major AI issues, Amazon employees recently questioned executives about their company's facial recognition products. Recently, Googlers have asked the AI Review Committee, which includes the president of the Heritage Foundation, On the other hand, he pointed to the history of lobbying for LGBT rights issues, and the company soon disbanded the board completely
"The diversity crisis in AI is well documented and extensive," the researchers concluded. It reflects the unequal workplace of industry and academia, imbalance of employment and promotion, stereotype of prejudice Can be seen in the re-packaging of biological determinism, artificial intelligence technology, automated systems and amplified. "