A new report from Common Sense Media is raising major red flags for parents as two popular AI tools, ChatGPT’s updated chatbot and OpenAI’s new video-creation app, Sora. While ChatGPT is still rated “high risk” for children, the group’s latest assessment goes even further with Sora, calling the app “unacceptable” due to the potential dangers it poses to kids.
Most parents have at least heard of ChatGPT, but Sora is newer to the scene. After a limited rollout earlier this year, Sora 2 landed in the iOS App Store in September, giving users the ability to create extremely realistic AI-generated videos with just a text prompt and almost no built-in oversight. That lack of guardrails is what experts say makes Sora especially alarming.
“It’s like ChatGPT, but for video instead of text,” explains Robbie Torney, senior director of AI Programs at Common Sense Media in a recent interview with Parents.com. “Users can create completely original videos and scroll through a feed of AI-generated content that looks and functions a lot like TikTok, except almost none of it is real.”
While Sora includes watermarks on its videos, those labels can be removed or easily missed once clips start circulating on other platforms like TikTok, Instagram, YouTube, or Discord.
“This makes it nearly impossible for anyone to tell what’s real anymore when AI-generated videos mix with authentic content across social media,” Torney says.
Common Sense Media’s assessment cites several major concerns as to why Sora is dangerous for kids. These concerns include:
- Little to no content moderation
- Graphic or harmful videos presented in upbeat, kid-friendly formats
- Extremely limited parental controls
- Minimal warnings on dangerous content
- A cameo feature that can recreate a child’s face and voice for anyone to use
Sora also has weaker safety systems than ChatGPT, according to Torney. Parents only have access to three controls, feed personalization, continuous scrolling, and direct messages, none of which allow them to monitor what kids view, create, or share.
Additionally, Sora “… easily generates eating disorder content, self-harm references, and dangerous activities that ChatGPT blocks,” Torney says, noting that the app offers very few crisis resources in response. One of Sora’s most unsettling capabilities allows users to upload their face and voice so the app can create a full AI version of them.
“The cameo feature lets users upload their face and voice to create an AI version of themselves that can star in AI-generated videos,” Torney explains. While designed to be playful, it opens the door to major misuse.
Once a cameo leaves the app, it can spread instantly and kids have little control over how their likeness gets used.
“Cameos can spread very quickly before a child has a chance to clamp down on the permissions that he or she has already given, and the child’s likeness can be used in humiliating or sexual content which can spread like wildfire throughout the internet,” says Yaron Litwin, CMO of Canopy.
ChatGPT isn’t without risks, but experts agree it has stronger guardrails than Sora. Still, Common Sense Media rated it “high risk,” largely because so many teens use it for companionship or emotional support, something the tool simply isn’t designed for.
“It’s designed to keep conversations going, not end them, even in mental health conversations where the goal should be rapid handoff to human care,” Torney explains. For ChatGPT, experts say usage may be okay with parental controls in place and clear boundaries about what it can and cannot be used for, especially when it comes to mental health.
“ChatGPT can be used more safely than Sora, but like all tech, it requires active parenting, not just turning on controls and hoping for the best,” says Torney.
As for Sora, “Given the significant risks and minimal protections, we recommend teens not use Sora,” Torney says plainly.
Parents can block the app for younger kids, but complete bans often don’t work for tech-savvy teens. That’s why many experts emphasize conversation over restriction.“Instead of focusing only on blocking or monitoring, I encourage huddling,” says Laura Tierney, founder of The Social Institute. “That means keeping open conversations with your child about what’s safe, what’s not, and how to make smart choices online.”