Warning labels about depression and anxiety on social media could reduce how much teens use the platforms, according to a Stanford Medicine study published Friday, Sept. 4, in JAMA Health Forum.

The finding lands less than four months before California's social media warning law takes effect Jan. 1, 2027. That law requires platforms to display health warnings to users under 17 on first daily access, again after three hours of active use and once per hour after that.

Researchers led by Anna Grummon, assistant professor of pediatrics at Stanford Medicine, and Thomas Robinson, a professor of child health at Stanford Medicine, tested 14 warning messages on 1,012 young people. About half were teens ages 13 to 17; the other half were young adults ages 18 to 29. Participants saw messages formatted as pop-ups on a simulated Instagram page.

Warnings about overall mental health harms and those specific to depression and anxiety resonated most. Participants rated every health warning as more effective than a control message about seatbelt safety. All but an addiction-focused warning outperformed a TikTok-style screen-time break prompt.

The study has limits. It measured whether participants believed the warnings would reduce their use, not whether their behavior actually changed. The researchers said their next step is testing whether the labels reduce social media use in practice.

"We haven't known what warnings for social media should say, or if they're promising for changing behavior," Grummon said. "Sometimes, subtle changes to warnings really matter."

Teens and young adults responded similarly to the warnings. Robinson said the researchers had expected the age groups to differ.

The study was inspired by a 2024 call from then-U.S. Surgeon General Vivek Murthy to add health warning labels to social media, modeled on tobacco and alcohol warnings. Four states have since passed such laws: California, Colorado, Minnesota and New York. A bipartisan coalition of 42 state attorneys general has petitioned Congress for federal action.

California's law, Assembly Bill 56 (AB 56), is among the most specific. It requires a "black box warning" covering at least one-fourth of a user's screen, displayed for at least 10 seconds. After three hours of use, the warning expands to 75 percent of the screen and stays for 30 seconds with no option to click through.

The mandated text reads: "The Surgeon General has warned that while social media may have benefits for some young users, social media is associated with significant mental health harms and has not been proven safe for young users." The Stanford team tested that California-specific language and found it performed on par with shorter health warnings, despite being longer and more complex.

Robinson said all state warning-label laws face legal challenges from the social media industry. A federal court in November 2025 temporarily halted Colorado's law on First Amendment grounds in a case brought by tech industry group NetChoice. California's law has not been blocked as of publication.

The pressures the study addresses are familiar to Palo Alto students. Paly sophomore Ethan Bradley told The Campanile in a July feature about reducing screen time that social media "takes over every part of your life" and makes him "a lot less productive."

The researchers' next milestone is a longer-term experiment measuring whether warnings change actual scrolling habits, not just intentions.