Thank you for inviting me today.

You have probably had the opportunity to hear from a lot of people about the harms of social media, so I will not repeat the laundry list again. Instead, I'd like to focus on two topics that hopefully will give context to that testimony and provide urgency for action.

First, I want to emphasize that we are profoundly underestimating the severity of social media's impact on children, due to limitations in how we observe and measure these effects. When researchers and policy-makers discuss the harmful effects of social media, they typically point to studies of teenagers documenting rates of self-harm, eating disorders and declining mental health among 16-year-olds, but these studies are echoes of the past, capturing the aftermath of social media exposure that began years earlier, typically around 12 or 13.

What's alarming is that when we talked to today's 12-year-olds and 13-year-olds, we discovered that they started on social media around age eight or nine. In 2022, 30% of American children between the ages of seven and nine were already active on social media platforms. That number is probably higher today. This creates what I call the telescope effect of our understanding of social media's impact. Like astronomers observing distant galaxies, we're always looking at information about the past: how social platforms were designed in the past, past usage patterns. This may be okay when we look at the stars, because the heavens change slowly, but when we examine the digital lives of teenagers, their rapidly changing world means that we end up continually surprised that rates of harm keep going up.

A seven-year-old is influenced and impacted in a meaningfully different way than a 13-year-old is. The children starting social media use today are doing so at ever younger ages, during even more crucial developmental periods and with even more sophisticated and engaging platforms than today's teenagers whom we're currently studying. If we don't act, we're on track to wake up in 10 years to realize that we've fundamentally altered a generation's development in ways that we failed to anticipate or prevent.

My second point concerns the emerging and under-reported threat of the rise of AI avatars and their impact on children's social development. These AI avatars are sophisticated virtual companions. They use artificial intelligence to engage in conversation, respond to emotions and build what feel like genuine relationships with users. They're designed to be always available, eternally patient and perfectly attuned to their users' interests and needs.

The leading provider of these AI avatars proudly announces that the average user—predominantly children under 18—spends two hours daily interacting with these virtual companions. This statistic should alarm us. Learning to navigate real human relationships is inherently challenging and sometimes uncomfortable. It requires compromise, patience and the ability to engage with others' interests and needs, not just our own. AI avatars, in contrast, offer a path of least resistance. They never disagree uncomfortably. They never have conflicting needs, and they never require the complex emotional labour that real friendships demand.

We need to expand our understanding of what constitutes social media. These AI-driven spaces represent a new frontier of potential harm, where the artificial ease of virtual relationships further erodes children's ability and motivation to build genuine human connections. If we don't act now to understand and regulate these technologies, we risk being blindsided by their effects, just as we were with social media platforms.

In conclusion, the problems we're seeing with social media are reflections of broader societal issues. The adults most negatively impacted by social media are often those already marginalized in our society, whether geographically, physically or economically. While in-person socialization carries real costs in terms of transportation, activities and time, online socialization appears free at first. The true cost is paid in terms of mental health, development and human connection.

Similarly, and perhaps most critically, the children most likely to become deeply enmeshed in these virtual worlds, whether traditional social media or AI-driven spaces, are likely to be our most vulnerable and marginalized youth. These are often the children with fewer opportunities for in-person social interaction, fewer resources for supervised activities and fewer adult mentors to guide them through the challenges of growing up or to provide context and support when they face online harm.

We must act now to ensure that children have appropriate and safe digital spaces, because their ability to meaningfully build relationships and connect will shape the world we all live in for decades to come.

Thank you.