AI’s social justice problem: It can amplify human bias – CNET [CNET]

Artificial intelligence now plays a key role in deciding who gets jobs, who gets into colleges, who gets loans, who gets accused of crimes and so much more. But recent work from researchers has shown that the algorithms driving AI are inheriting — and in some cases even amplifying — the biases behind the inequalities and injustices in our society, especially for women and people of color. 

A documentary called Coded Bias just landed on Netflix this month and it tackles this issue head on. Director Shalini Kantayya joined CNET’s Now What series to talk about the film.

Read more: Coded Bias review: Eye-opening Netflix doc faces racist technology

Kantayya talked about how the documentary opened her eyes to one of the biggest challenges society is facing as we move into the AI age.

“I hadn’t realized the extent to which algorithms, machine learning and AI are increasing becoming the gatekeeper of opportunity — deciding such important things as who gets hired, who gets what quality of healthcare, even who gets the (COVID-19) vaccine, or how long a prison sentence someone might serve,” Kantayya said. 

“So as I started to understand the extent to which we are outsourcing our decision-making to machines … I began to realize that these same systems that we’re trusting so implicitly with decisions that are essentially changing human destiny have not been vetted for racial bias or gender bias or — or more broadly that they won’t hurt people or have unintended consequences and cause harm.”

We also talked to Kantaya about the conflict between the Silicon Valley narrative of AI and the Hollywood narrative of AI, which Princeton professor Ruha Benjamin brought up in her appearance on Now What.

Watch the whole interview with Kantayya and then catch the documentary on Netflix.