Building a framework to better protect teens on social media

As debate rages over the best way to protect young users from the harms of social media, here’s a review of the key issues in question, and the currently proposed solutions to address these problems, and ensure more safety for teen users.

Are teen social media bans necessary?

A growing body of academic research indicates that social media can be harmful to young users, exacerbating issues relating to negative self-perception, misinformation, exposure to online criticism and bullying, potential exposure to predators, and more.

Yet, at the same time, counter research indicates that social media engagement is actually beneficial for teens, as it facilitates connection, and has become a critical element of the modern interactive process.

Though the actual data has largely been separated from the debate either way, because according to polling, the majority of voting adults actually want action taken to restrict teen social media usage.

Given the support for action, it seems that teen social media bans will be enacted in many regions.

Can teen social media bans be effectively enacted?

The key question here relates to the viability of user age checking, and the capacity for social platforms to keep teen users out of their apps

So far, the signs on this element are not great. In Australia, which was the first country to implement a large-scale social media ban for people under 16, the vast majority of underage users are still accessing restricted apps according to the Australian eSafety Commissioner’s latest report, which covers the first six months of the ban.

In Australia’s case, regulators tested a range of age-checking processes, including age verification through government ID, facial recognition for age estimation and age inference based on activity. It determined that there are many services available which can accurately determine user age, and offer a means for platforms to abide by the rules, though it didn’t mandate a specific approach that all platforms should take.

Given the data, that approach is not working, and without this key first step, the rest is irrelevant.

Regulators need to establish a more structured, uniform approach to age checking that all platforms are required to abide by. In EU, they are testing a form of government ID for teens, while Meta claims that AI can largely solve the problem, while also suggesting that app stores should be responsible for age-checking.

Whatever the approach, this requires a definitive approach to ensure compliance.

Should messaging apps also be restricted?

It’s also a relevant note that, at this stage, most of the proposals do not include messaging apps in scope for age restrictions.

That’s relevant because in recent times, more and more social sharing has migrated to private messaging groups. Indeed, in the Australian 6-month update on its teen social media ban, it also noted that young teens were spending more time in messaging apps and playing online video games as a result of the change.

The counterargument is that messaging apps are now a utility, and cutting them off could restrict the capacity for teens to get in touch with parents or friends, which could pose a safety risk in itself. But it’s another consideration, that messaging platforms may become a proxy for social sharing, limiting the effectiveness of any restrictions.

Should algorithmic amplification be removed for teen users?

Last week, the Australian government proposed an alternative approach to teen safety online, which could see the implementation of restrictions on the use of algorithms for teen users specifically.

Which could be a good move. Research clearly indicates that algorithmic amplification can be harmful, given the incentives created by algorithmic boosting of certain types of content.

The balance of academic research suggests that anger is fueled by algorithmic amplification, whether intentional or not, which then incentivizes more divisive takes from creators, in order to drive more response.

Banning algorithmic amplification, then, could reduce that incentive, and may lead to a more controllable, less harmful experience for teens.

Worth noting, also, that Meta recently included a provision that teen users in the U.S. will be able to choose a non-algorithmic feed as their default as part of its $18 billion settlement deal in a hearing over social media harms.

Should AI chatbot usage also be restricted?

A rising area of debate relates to the use of artificial intelligence-powered chatbots, and whether these could also be causing harm for teens.

The actual valuable use cases for AI chatbot usage among teens are less convincing than the reasons for restricting them, with AI tools causing all kinds of concerns for educators, and leading to false depictions that can be used for cyberbullying, while also driving misinformation and misunderstanding due to confirmation bias that’s inherent in AI tools.

Last week, the State of New York announced a moratorium on student-facing generative AI use in schools for children in 2-K through to the 8th grade, while in Norway, AI usage has been banned for elementary school pupils, and restricted for older children, in order to prevent a ‌negative impact on learning, as reported by Reuters.  

Given the broader focus of teen social media restrictions, this should also be under consideration.

Will restricting social media usage among teens help?

As part of its recent $18 billion settlement agreement, Meta agreed to implement a range of restrictions on teens, in order to address concerns around potential harms.

Those restrictions are:

These seem like positive, logical steps, which were actually proposed by a New Mexico judge in another recent court case about Meta’s teen protection efforts.

These should be implemented, though again, without accurate age verification, they won’t be viable either way.

Which nations are considering teen social media bans?