Civil society organizations are demanding action against ‘sexist’ AI disinformation

Global civil society groups say women and sexual minorities are disproportionately affected by non-consensual AI content. – Copyright AFP Stefani REYNOLDS

Anuj CHOPRA

More than 20 international civil society organizations will call on major tech companies to strengthen their AI policies to combat “sexist and misogynistic” misinformation plaguing social media platforms, according to the draft of an open letter seen by AFP on Thursday.

The letter to the CEOs of six giants – Meta,

“It is clear that this harm is not being felt equally,” said the letter, signed by 27 digital and human rights organizations, including UltraViolet, GLAAD, the National Organization for Women and MyOwn Image.

“In particular, women, transgender and non-binary people are uniquely at risk of experiencing negative consequences from AI-based content on social media.”

The letter, which the groups say will be made public on Friday, makes a dozen recommendations to strengthen AI policy.

These include clearly defining the consequences for posting non-consensual explicit material – which should include suspension of repeat offenders – implementing a third-party tool to detect AI-generated images, and clearly labeling such contents.

The groups also demanded a coherent process for users to flag and report harmful content and that platforms conduct comprehensive annual audits of their AI policies.

– Misogynistic, sexist attacks –

The letter comes barely a month before what is widely billed as America’s first AI election on November 5. The tight race for the White House has produced a flood of misinformation.

A particular target of gendered misinformation is Democratic Party candidate Kamala Harris, who has included a barrage of misogynistic and sexist stories attacking the first Black, South Asian and female vice president in American history.

“These harms silence us online, violate our right to control our own image, and distort our elections,” said Jenna Sherman, campaign director at UltraViolet.

“But worse, they normalize and even algorithmically codify sexual exploitation and reinforce harmful stereotypes about gender, sexuality and consent.”

The proliferation of non-consensual deepfakes is outpacing efforts to regulate the technology worldwide, experts say, with several photo apps digitally undressing women and manipulated images fueling ‘sextortion’ rackets.

While celebrities like singer Taylor Swift and actress Emma Watson have fallen victim to deepfake porn, experts say women who are not in the public eye are just as vulnerable.

“AI technologies have further facilitated the creation and spread of gender-based harassment and abuse online,” said Ellen Jacobs, senior US digital policy manager at the Institute for Strategic Dialogue, one of the organizations that signed the letter.

“We need effective policies that specifically address the increased risks for women, girls and LGBTQ+ people.”

The platforms did not immediately respond to a request for comment prior to the letter’s publication.

“The world’s largest platforms have shown that they are not equipped to handle the rise of AI-facilitated hate, harassment and disinformation campaigns, including deepfakes and bots that can widely spread hate-based imagery,” said Leanna Garfield, social media safety program manager at GLAAD.

The platforms “must take concrete action now so that everyone can feel safe online.”