Online safety watchdog says tech giants failing to detect child abuse
Julie Inman Grant is the head of Australia’s online safety regulator.(AAP: Lukas Coach)
In short:
Major tech companies including Apple, Google, Meta and Microsoft are still failing to do enough to detect and stop child sexual abuse, eSafety has claimed.
A new report highlights gaps in proactively preventing exploitation and sexual extortion activity on livestreaming, video and messaging platforms.
What’s next?
ESafety Commissioner Julie Inman Grant is calling on the tech companies to lift their efforts to prevent the spread of exploitation material.
Tech giants are still failing to detect and stop child sexual abuse on their platforms despite having the technology and resources to tackle it, a new report has found.
Australia’s eSafety Commissioner has identified major gaps in how some of the world’s biggest companies identify and prevent online child exploitation, including livestreamed abuse, grooming, AI-generated material and sexual extortion.
A new transparency report released today has acknowledged some “welcome improvements,” but also highlighted how the eight companies monitored — Apple, Discord, Google, Meta, Microsoft, Skype, Snap and WhatsApp — are not consistently applying “basic” technology to protect children.
Social media platforms were required to provide data to eSafety as part of the laws introduced in December.(ABC News: Abbey Haberecht)
ESafety Commissioner Julie Inman Grant said combating child sexual abuse online was not just an issue of legality, but a critical matter of “corporate conscience and accountability” for the tech giants.
“These companies are no strangers to innovation and have the technological capability to develop new technologies to detect this harm in the fight to end live online abuse of children,” she said.
“This represents a failure of industry will, not of technical capability.”
‘Inadequate’ detection of child sexual exploitation in video calling services
The report found companies are not consistently applying tools available to detect when abuse is occurring on live video.
For example, Meta does use detection technology for Facebook and Instagram live, but not its Messenger platform.
Similarly, Google does not use the tools for Google Meet, but does for YouTube.
Other providers of video calling services not using proactive detection tools on any of their services included Apple FaceTime, Discord’s live and video call functions, Microsoft Teams, Snapchat and WhatsApp.
Ms Inman Grant said the companies have had years to make their spaces safer, but continued to fail at this responsibility.
“Here we’ve highlighted eight of the largest and very mature trillion-dollar companies, and they still don’t have their house in order,” she said.
“These are high-risk, high-impact issues that have devastating consequences for young Australians.”
Young Australian men the biggest targets of sexual extortion
eSafety also found Apple, Discord, Google’s chat, meet and messages, Microsoft Teams, Skype and Snap were not using any language analysis technology to proactively detect sexual extortion of children or adults on their services.
By comparison, Meta’s Facebook and Threads, Microsoft’s Xbox and Google’s YouTube used these tools on all material by default.
The eSafety Commission found that platforms are not using appropriate technology to proactively detect sexual extortion on their services.(Graphic: Unsplash: Steinar Engeland/ABC News: Sharon Gordon)
Sexual extortion is a form of blackmail where someone threatens to share a nude or sexual image or video unless the person targeted gives into their demands.
Ms Inman Grant said eSafety was “overwhelmed” with reports of organised criminal networks infiltrating mainstream social media and messaging platforms “targeting young Australian men”.
“This has resulted in suicides, so it’s a pretty heinous form of sexual exploitation, and they (the criminals) are doing it at scale,” she said.
Ms Inman Grant said eSafety had sent intelligence reports to tech companies with lists of “language indicators” to help them detect extortion, but few had used the “basic AI” needed to do this.
“These companies aren’t eating their own dog food,”
she said.
“They’re developing high end AI tools and services for customers, but they’re not using it on their own platforms.
“That kind of beggars belief.”
Some ‘welcome’ improvements
Microsoft was acknowledged for expanding its use of tools to detect known child abuse material in OneDrive and email attachments in Outlook sent worldwide.
Meanwhile, Snap has reduced the time taken for moderators to reach an outcome after receiving a report of abuse material from 90 minutes to just 11 minutes.
Google has launched a feature that blurs incoming images containing nudity before viewing, as has Apple for accounts with users who have declared themselves to be aged under 13.
The eSafety report is the second of four to be published in the coming months.
It is based on disclosures the eight companies are legally required to provide, or face fines of up to $825,000 for every day they do not respond.
Ms Inman Grant said reports were a “reputational lever” that could be more effective than regulatory crackdowns in getting companies to do more than the “bare minimum” on safety.
“Without transparency, you can’t have any accountability,” she said.
“We just need to keep plugging away and keep exposing where the holes are, you know if you can do it on one platform … if you can deploy language detection technology or newer technologies … on one service, you can do it on another,” she said.
Loading
