How Pornhub’s New AI Identifies Adult Stars in Videos

Have you ever wondered why you can search videos by combining interests? It’s because countless people manually classify and tag video content, but artificial intelligence is now taking over that role.

Pornhub announced a new AI model that leverages computer‑vision technology to automatically detect and identify the names of adult performers.

The algorithm, after training, can recognize more than 10,000 adult stars by scanning video frames and understanding visual cues. Pornhub fed the model thousands of videos and official photos of performers so it could learn to return accurate names.

To reduce errors, the site asks users to confirm whether the AI‑generated tags and classifications are appropriate, allowing them to support or oppose the results, which helps the algorithm become smarter.

"Now users can find specific adult stars based on their preferences, and we can provide the most precise search results," said Pornhub Vice President Corey Price. "Our model will play a key role as the platform adds over 10,000 new videos daily."

In testing, the model scans about 50,000 video segments each day, automatically adding or removing tags.

The system also distinguishes content categories—for example, videos labeled "Public" are outdoor shoots, while those under "Blonde" must feature a blonde performer.

The company plans to catalog all 5 million videos next year with AI assistance, making it easier for users to find the clips that best match their expectations.

Earlier research has used computer‑vision algorithms to describe adult content, and a developer previously built a bot with Microsoft AI that could continuously observe and interpret such material.

Pornhub’s advancement pushes this concept further, posing a nightmare for human video reviewers worldwide.

While AI’s involvement may seem astonishing, the adult industry’s relentless focus on search‑engine optimization makes such technology a natural evolution.

Adult content services have a broad market and appeal to all ages, forming a significant revenue component for these companies.

Human taggers, who have long painstakingly classified videos, may soon become victims of automation, but they also gain the opportunity to let automation work for them.

Compiled by Linux China / Dongfeng Weixiao Link: https://linux.cn/article-8964-1.html