Google has admitted to a loophole that allowed Clive Palmer’s United Australia Party to spread misinformation multiple times on YouTube without being banned for life under a three-strike policy.
During a tense hearing before a parliamentary inquiry into social media and online safety, Google’s director of government affairs and public policy Lucinda Longcroft was grilled about the proliferation of misinformation on YouTube – which is owned by the tech giant.
She said Covid-19 misinformation policies needed to be “robust” and the company took actions against repeat offenders by removing their accounts for life if they received three strikes over 90 days.
“When a first strike is issued we remove privileges of a channel owner for a full week,” she said.
“If a second strike is received in a 90-day period, we take down privileges for two weeks.
“If there is a third strike that information and channel is terminated perpetually.”
But Labor’s Tim Watts said he had reported nine UAP videos for Covid-19 misinformation.
“You took down six but we don’t get to three strikes from doing it?” he asked.
“If a number of videos are found to violate at the same time they are bundled into one strike,” Ms Longcroft answered.
“I’ll remember to stretch them out next time,” Mr Watts said.
Mr Watts also pointed to a UAP advertisement on YouTube that used incomplete extracts from the Therapeutic Goods Administration about vaccine adverse events that was deemed to be seriously misleading.
“That was removed eventually but not before the UAP had spent somewhere between $100,000 promoting it in advertising,” he said.
Ms Longcroft said Google had “special election transparency policies” to deal with vulnerable areas like elections, which had even stricter guidelines over who can advertise on its platforms.
“While of course we do not and are not comfortable with any mistakes that are made, we are rectifying those mistakes, and taking all available action,” she said.
Liberal committee chair Lucy Wicks later put a series of examples to Ms Longcroft to see whether Google would remove them for breaching standards, but was not given a direct answer, instead being told “context” was relevant.
“Would this comment break your community standards for calling a man or woman a whiny little bitch?” she asked, in reference to YouTuber Friendlyjordies labelling retiring backbencher Nicole Flint that in a podcast.
“Those standards would have to take into account the context of the person, the nature of the person who had made the comment,” Ms Longcroft said.
“I’ll ask you one with language that indicates someone should kills themselves or (referencing a specific method of suicide). Would that be removed if requested?” Ms Wicks continued.
“We have a process by which our algorithms are designed to remove violative content and also to reduce content that while it doesn’t violate our policies, is at the broader line of those policies,” Mr Longcroft answered.
“A comment about a woman ‘I bet she rages so hard a natural disaster occurs each time she has her period’?” Ms Wicks later asked.
“These comments of course are distressing but the context in which they are made is taken into account whether Australians and users all over the world should have access to them and to the diverse range of content that is essential to support our democracy,” Ms Wicks answered.
