Home
Elementary School Student Draws a Mustache to Fool AI Age Verification, Leaving a Silicon Valley Engineer Silent
Amidst tightening global regulations, leading social networks have implemented rigorous age verification protocols to restrict minors from accessing specific features. To safeguard user privacy, many platforms have embraced an innovative device-side facial age estimation system. Yet, ironically, this advanced technology has recently been effortlessly circumvented by online "amateur experts" using remarkably low-tech tactics.

Basic Drawings and Fake Facial Hair Succeed
In a lighthearted test of AI capabilities, a 12-year-old boy drew a mustache above his lip with an eyebrow pencil and was accurately estimated to be 15 years old, successfully passing verification. Even more amusingly, some users simply drew two dots and a line on their thumb to represent eyes and a mouth. By moving their head side-to-side as prompted, the AI determined the subject was between 13 and 15 years old.
Technical Weaknesses of Lightweight Models
This embarrassing vulnerability stems from the fact that these age estimation models operate entirely on the user's device. Due to limited computing power, the models cannot be overly complex. AI primarily relies on visual cues like eye shape and skin texture for approximate inference. Furthermore, to avoid incorrectly blocking legitimate users during the initial rollout, platforms set the judgment threshold relatively leniently, creating an opening for carefully crafted deception.
Related article
AI startups accelerate revenue growth
As established firms and emerging ventures scramble to leverage artificial intelligence, numerous AI startups report that their revenue is not merely expanding, but accelerating rapidly, achieving subsequent milestones in increasingly shorter periods
Tencent Hunyuan integrates critical point size theory into PPO reinforcement learning, boosting throughput 2.29x and cutting GRPO training time by 29 percent
As reinforcement learning for large models scales up to larger GPU clusters and denser training datasets, training efficiency has emerged as a critical priority. Recent research from the Tencent Huan Yuan team addresses a frequently overlooked challe
Replit’s Amjad Masad on the Cursor deal, fighting Apple, and why he’d rather not sell
Amjad Masad has spent the last decade building Replit, yet the past 18 months have been a transformative period. The AI coding platform has surged from $2.8 million in annual revenue in 2024 to a trajectory toward what Masad describes as a billion-do
Related Special Topic Recommendations
Comments (0)
0/500
Amidst tightening global regulations, leading social networks have implemented rigorous age verification protocols to restrict minors from accessing specific features. To safeguard user privacy, many platforms have embraced an innovative device-side facial age estimation system. Yet, ironically, this advanced technology has recently been effortlessly circumvented by online "amateur experts" using remarkably low-tech tactics.

Basic Drawings and Fake Facial Hair Succeed
In a lighthearted test of AI capabilities, a 12-year-old boy drew a mustache above his lip with an eyebrow pencil and was accurately estimated to be 15 years old, successfully passing verification. Even more amusingly, some users simply drew two dots and a line on their thumb to represent eyes and a mouth. By moving their head side-to-side as prompted, the AI determined the subject was between 13 and 15 years old.
Technical Weaknesses of Lightweight Models
This embarrassing vulnerability stems from the fact that these age estimation models operate entirely on the user's device. Due to limited computing power, the models cannot be overly complex. AI primarily relies on visual cues like eye shape and skin texture for approximate inference. Furthermore, to avoid incorrectly blocking legitimate users during the initial rollout, platforms set the judgment threshold relatively leniently, creating an opening for carefully crafted deception.
AI startups accelerate revenue growth
As established firms and emerging ventures scramble to leverage artificial intelligence, numerous AI startups report that their revenue is not merely expanding, but accelerating rapidly, achieving subsequent milestones in increasingly shorter periods
Tencent Hunyuan integrates critical point size theory into PPO reinforcement learning, boosting throughput 2.29x and cutting GRPO training time by 29 percent
As reinforcement learning for large models scales up to larger GPU clusters and denser training datasets, training efficiency has emerged as a critical priority. Recent research from the Tencent Huan Yuan team addresses a frequently overlooked challe
Replit’s Amjad Masad on the Cursor deal, fighting Apple, and why he’d rather not sell
Amjad Masad has spent the last decade building Replit, yet the past 18 months have been a transformative period. The AI coding platform has surged from $2.8 million in annual revenue in 2024 to a trajectory toward what Masad describes as a billion-do











