The idea of opting out of algorithms sounds great, but raises the question: what would you get instead?
So a more powerful change the government could have chosen instead would be to make it so you get a chronological feed by default; to get algorithmic recommended content you’d have to opt in.
The draft legislation puts the onus on social media platforms to identify and manage risks around content, such as videos promoting disordered eating. But it’s hard to know exactly how.
So, if you opted out of the algorithm in Instagram, which of those three would be impacted?
What’s been proposed?
Take Instagram, for example. There’s the main feed, which is one recommender algorithm. There’s disappearing stories, which is a different algorithm. And then there’s reels, which is a third algorithm. They work in different ways and do very different jobs.
- users could opt out of social media recommender algorithms via pop-up messages
- social media companies would be expected to identify and manage risks on their platforms
- the onus would be on platforms to stop harmful content (such as illegal pornography) from being shared.
As with the social media ban, it looks like the government will try to put legislation in place before sorting out the technical details of how it will work. That’s a mixed blessing.
Under the proposed digital duty of care legislation:
The ABC reported that, for under-18s:
The proposed digital duty of care legislation would hold social media companies to greater account for content on their platforms. Opposition leader Angus Taylor said the Coalition had yet to see the proposal but warned against efforts to “censor social media”.
Will it help?
That said, the rest of the world has shown a much stronger appetite for regulation of content on platforms in recent years.
What matters most with the digital duty of care is ensuring social media platforms are forced to be transparent about their compliance.
So what’s been proposed in these digital duty of care laws? And will it help reduce harm?
The devil will be in the detail, which we won’t know until the full plans are released in the week ahead.
With the proposed digital duty of care laws, it looks like there might be provisions for the eSafety Commissioner and for researchers to be in involved in finding out about and testing what companies are doing to ensure compliance with these laws.
The digital duty of care plan is separate to the government’s proposed Online Safety Amendment (Strengthening Enforcement for the Social Media Minimum Age) Bill, which aims to toughen existing rules around the under-16 social media ban.
Opting out of algorithmic content
about half a dozen additional areas of ‘psycho-social harm’ would be identified, including content about body image and bullying.
One thing Australia got wrong with the social media ban is that we asked the companies to comply with the laws but they don’t have to tell us exactly how they did it.
That said, there have long been calls for digital duty of care laws in Australia – since well before the under-16s social media ban – and it’s good to see these calls being considered.
It’s also unclear how parents might manage all this opting-out. If it comes down to vast and complex parental control dashboards for each platform and online service, that places a large burden back on parents.
We haven’t seen the detail but it sounds like this aspect of regulation is being taken seriously – and that’s very promising.
Fines of around 0 million for breaches of the proposed digital duty of care laws sound impressive, but it’s a tiny fraction of what some of these tech companies make annually. Making fines a percentage of a company’s yearly revenue in Australia makes more sense.
It’s easy to see how moves to more strongly regulate this could lead to an angry midnight phone call from US President Donald Trump to Australian Prime Minister Anthony Albanese.
Harmful content harder to regulate
Read more:
Australians could soon be able to choose what appears in their social media feed. How would this work?
Under the plan, breaches of the law would lead to fines of more than A0 million.
Historically, holding tech companies responsible for the content shared on their platforms has been very difficult. In the American tradition, the idea is that users (not companies) are responsible for such content.
It’s been so long since people have seen a chronological feed (as opposed to one serving you algorithmic content tailored to you), it would feel very weird at first.
A promising start
The onus should be on platforms and other online entities to better test the risks of their products, before features are rolled out.
Social media users could switch off recommender algorithms in their social media feeds, under new draft laws the Australian government plans to unveil this week.
It means we are not beholden to companies saying, “This isn’t possible”. But it also might make implementation and enforcement tricky.
In an ideal world, that would mean they could mitigate the risks and evaluate whether features or tools do more harm than good before they are released. If they are found to being causing harm, platforms could not roll them out at all, or ensure they aren’t available to children and young people.
My suspicion is that the vast majority of people won’t elect to opt out of any algorithms, and will stick with the recommended content in general.
