‘My Feed, My Way’ laws will give your algorithm an off switch
By Ben Kaldi, the Australian Science Media Centre
The Australian Government has proposed new laws that would force social media platforms such as Meta to offer an opt-out option for algorithmic feeds, which would allow users to choose to have their default feed include only personalised content.
Additionally, social media providers will be legally required to protect people under the age of 18 from harmful content, including content that relates to eating disorders, misogyny, pornography, crime, dangerous stunts, abuse and bullying.
Earlier this week, the AusSMC chatted with several researchers to discuss the likely outcomes for Australians if the laws are initiated.
“The Australian Government’s ‘My Feed, My Way’ represents an important next step in protecting all users from many of the harms associated with social media," Dr Elizabeth Englezos, a lecturer from the Law School at Griffith University, told the SMC.
“Up until now, social media companies have escaped liability for harmful content on free speech grounds. However, the recent legal claims focusing on harmful design — such as infinite scroll and algorithmically-mediated feeds — have sidestepped the traditional defences of Free Speech or immunity under the Communications Decency Act.”
Many experts suggested these new laws will likely have benefits, explaining that past research has repeatedly shown that the algorithms amplify the harms we may already be seeing on social media.
“The recommender has one job - maximise engagement, and what holds attention best is what provokes. Addictive design and the documented damage to young people's mental health both follow from that objective,” said Professor Marian-Andrei Rizoiu of the University of Technology, Sydney.
Dr Natasha Anne Rappa from Murdoch University said the initiative will raise awareness of the recommendation system underlying many social media platforms and its tendency to fixate on negative events and stereotypes, agitate users, and ultimately push them towards holding extreme positions.
In addition to rage-bait, the algorithm has a tendency not to discern between good information on health and misinformation.
“Our research has shown that health information targeted at healthy individuals is highly misleading and that substantial financial conflicts of interest exist, " said Dr Brook Nickel of the University of Sydney’s Health Literacy Lab.
"Misleading health claims can also spread quickly, particularly when they’re sensational, fear-based and/or explicitly designed to sell you something. The public shouldn’t have to become experts at spotting health misinformation just to safely navigate their social media feeds.”
She said that giving the public greater control over what algorithmic content they see is an important public health intervention.
Experts suggested that the laws will also need to make it as easy to opt-out of the algorithm as it is to opt-in, questioning any default setting.
“Opt-out is the wrong default," said Professor Rizoiu.
"People almost never change defaults, and the platforms know it. Algorithmic curation should be off until the user turns it on, and turning it on should require a short induction on what it does to you. Opt-in, not opt-out.”
The real test of these new laws, according to Dr Armin Alimardani from Western Sydney University, will be in measuring changes.
“Platforms must show that their safeguards reduce harmful exposure and damaging behavioural patterns over time. This would require third-party supervision and assessment. Regulators and independent researchers need meaningful access to data and auditing methods. Australia should judge compliance by measurable outcomes and not box-ticking processes,” he adds.
You can read the full Expert reaction here.
This article originally appeared in Science Deadline, a weekly newsletter from the AusSMC. You are free to republish this story, in full, with appropriate credit.
Contact: Ben Kaldi
Phone: +61 8 7120 8666
Email: info@smc.org.au