AI and social media bans won’t make California kids safer online. It may even be harmful

For more than two decades, my research group has been studying and supporting young people with disabilities, as they create and use digital technologies.

In one study of Autcraft, a Minecraft server built for autistic kids, a 13-year-old boy described a rich social life there that included friends he trusted, builds he led, other kids he comforted when they were struggling. 

And yet, when describing this rich social life, he asked, “Will I always be not social?”

This quote has stayed with me. This sweet, social child was measuring his life against a definition of “social” that was not built for him. And the way his parents, schoolteachers and other well-meaning adults were using that definition made him feel terrible about his own experiences and the things that made him happy.

Californians are right to worry about kids online. We are right to demand of our lawmakers, teachers, administrators and tech companies responsibility for what we are all collectively bringing into the world. The lawsuits behind this year’s chatbot bills involve real children who were harmed. 

Somehow, though, we keep behaving as though there is only a binary choice to either ban technologies, whether phones or AI, or keep the status quo. Both options fail the kids who need support the most.

Access bans sound protective, but the evidence so far mostly suggests they are not. 

Australia’s under-16 social media ban, evaluated after three months by both the government’s own regulator and a peer-reviewed medical journal study, found more than 80% of kids still using restricted platforms. They fake their ages, borrow logins and find ways around the ban.

Worse, even when bans do reduce usage, evidence indicates they don’t actually improve things we care about, like mental health and academic performance.

What concerns me most is the possibility that we are now undoing the decades we have spent destigmatizing the use of a variety of assistive technologies and participation in online communities.

We know that many kids go online because these are some of the few communities in which they truly feel comfortable, like the 13-year-old who used Autcraft as his primary social space. Research consistently shows LGBTQ youth find identity support and community online that isn’t always available at home or school. 

Autistic and other neurodivergent kids often find their best friendships in text-based or game-based spaces, where eye contact and improvised small talk aren’t the price of entry.

To be clear, the status quo is not the answer. There are substantial problems with these platforms, even if we restrict our thinking to adults. And none of them were built with childhood or disabilities in mind. 

Change needs to happen. It’s just that we should design better platforms for everyone, rather than banning or restricting them for some.

Recently, with the passage of Assembly Bill 1709, California lawmakers attempted a design-oriented framing, which I greatly appreciate. The law requires social media companies to stop using some of the more problematic design features, such as infinite scroll. 

The problem, for me and for groups like the Electronic Frontier Foundation, is that this law effectively bans social media for people under 16, given how broadly things like “addictive features” are defined. It also requires age verification, which largely doesn’t work and allows the same companies we are worried about to amass even more data about not only kids but all of us, with limited ability to mitigate these risks.

We already have examples of alternative, optimistic futures that don’t create a surveillance state. 

Autcraft is safe by design, with community-enforced rules, trained volunteer moderators and a culture built around acceptance. Likewise, other online communities that focus on popular media and games enable connection and learning that looks very different from what young people encounter in traditional schools. 

Rather than have their curriculum, ideas and social opportunities dictated to them with adult preferences and neurotypical norms, these “neurodiversity oases” allow youth to be in charge of learning and to be driven by an ethos of acceptance within a framework of passion for shared interests. 

The success of these platforms shows we don’t need to ban children from well-designed platforms. Good design can and does work. 

We should hold platforms accountable for what they design and encourage them to make better experiences for everyone, rather than banning some people while creating a surveillance infrastructure that risks harming us all.

California has a real choice this year: pass bans that push youth further from spaces they feel safe or start setting standards for platforms so those spaces actually are safe. The 13-year-old who wondered if he’d always be “not social” was asking us to see his social life, not to take it away.