eSafety report shows 'serious gaps' in how tech firms tackle harmful content
Make us a preferred source
Make us a preferred source on Google
The news: Australia's eSafety Commissioner has released a report containing "world-first insights" about how tech platforms are dealing with online proliferation of terrorist and violent extremist material and activity (TVE).
The numbers: The report — which summarised responses from Google, Meta, WhatsApp, X, Telegram and Reddit on steps each company is taking to tackle the harmful content and conduct — found that between April 2023 and February 2024:
- Google received 258 user reports about suspected AI generated synthetic TVE by its AI bot Gemini, and 86 user reports of suspected AI generated synthetic child sexual exploitation and abuse material by Gemini.
- WhatsApp took more than a day, Threads 2.5 days and Reddit 1.3 days to respond to user reports of TVE.
- Google only used hash-matching to detect exact matches of TVE content, even though technology which can detect varied versions of TVE is available. In the days following the 2019 Christchurch attack — where 51 people died following a livestreamed terror attack — Meta reported 800 different versions in circulation.
- Reddit and WhatsApp human moderators covered 13 and six languages respectively and only one of the top five languages, other than English, spoken in Australian homes, while Google covered approximately 80 languages and Meta 109 languages, including all top five languages (Arabic, Cantonese, Mandarin, Vietnamese and Punjabi). Telegram covered 47 languages but only two of the top five languages spoken in Australian homes, other than English.
The context: The eSafety report found that Facebook Messenger did not take steps to detect livestreamed TVE despite the use of another Meta product Facebook Live in the Christchurch attack. There was also no mechanism for users not logged-in to Facebook or YouTube to report livestreamed TVE.
Meanwhile, WhatsApp rolled out Channels, which is not end-to-end encrypted, without the use of hash-matching for known TVE and reported that only during the report period did it start working on its implementation.
What they said: "Ever since the 2019 Christchurch attack, we have been particularly concerned about the role of livestreaming, recommender systems and of course now AI, in producing, promoting and spreading this harmful content and activity," eSafety commissioner Julie Inman Grant said.
The source: eSafety Commissioner media release
