Submission + - OnDevice AI could stop abuse before it happens (linkedin.com) 1
bazorg writes: A recent interview and discussion on The Rest is Politics podcast episode 495, at about 29m30s, highlighted a striking claim from childprotection experts at the International Justice Mission: the technology already exists to prevent the creation and transmission of child sexual abuse material at the device level, but governments and manufacturers have not yet required it.
According to safetytech organisations, modern ondevice AI can reliably detect when a user attempts to create, view, or livestream illegal sexual content involving children—and block it before it’s produced or shared. Advocates describe this as “safety by design”: building phones, laptops, and operating systems that simply cannot be used to record or broadcast abuse.
The call to action is blunt: governments should require device manufacturers and OS vendors to implement ondevice protections, whether using existing tools or developing their own. Without regulation, companies have little incentive to deploy these capabilities despite their availability.
According to safetytech organisations, modern ondevice AI can reliably detect when a user attempts to create, view, or livestream illegal sexual content involving children—and block it before it’s produced or shared. Advocates describe this as “safety by design”: building phones, laptops, and operating systems that simply cannot be used to record or broadcast abuse.
The call to action is blunt: governments should require device manufacturers and OS vendors to implement ondevice protections, whether using existing tools or developing their own. Without regulation, companies have little incentive to deploy these capabilities despite their availability.
