Nuix confident that lack of diversity in AI can be remedied

Investigative analytics and intelligence software company Nuix is warning that the lack of diversity in AI is providing biased data relating to issues such as gender, race and culture.

Nuix Chief Technology Officer Alexis Rouch said she was concerned about what algorithms were being trained to do, as AI is only as good as the humans behind it and diversity is lacking from the coders behind the algorithms.

“Making machines think like humans requires having the right input, and one of the issues is that very few women are working in AI today and that can distort results. Globally just 22% of AI jobs are held by women, while in Australia only 7% of women have received training, compared to 17% of men,” she said.

Issues are emerging in a number of areas, including the processing of health data where most medical research is conducted on men — and the use of AI to interpret trends in health does not deliver accurate outcomes.

“Another example is that many companies use AI to filter job applications, and women whose resumes have career gaps because they have taken time off to have children can be eliminated from consideration for jobs when recruitment searches are done through AI,” Rouch said. “The lack of diversity is very widespread, and we are very concerned that the problem is amplifying and accelerating existing biases which could impact on our customers.”

Nuix says it has taken a strategic approach, thinking laterally to remedy the diversity issue by tapping into different talent pools and focusing on building diverse teams and responsible and ethical AI.

“We have purposely built our AI models using a diverse team of dedicated men and women from a wide range of backgrounds, including a schoolteacher, a philosophy major, a filmmaker and a full-time parent returning to the workforce,” Rouch said. “It can be done. I’ve got to almost 50/50 of women in my tech teams previously — it just requires a very focused and proactive approach in reaching out to different talent pools as part of your recruitment practices.”

Top image credit: iStock.com/gorodenkoff