🐙 In this inspiring animated video, Octo — our wise and witty octopus — invites you to become a Pharos Lighthouse Keeper, a community ambassador shining light across the Web3 ocean.@pharos_network
Help train better AI – and earn while you do it.
I'm contributing on @PerleLabs, where verified human feedback powers real AI data.
Join me here: https://t.co/YclEAGqOXX
Middle The middle image (the Leaning Tower of Pisa) is most likely AI-generated. The lighting and texture look slightly too uniform and polished, and the structure appears overly clean and symmetrical despite the real tower’s subtle irregularities and surface wear.
Alright… this is it. Final Human CAPTCHA challenge 👉👉
We saved the best for last. There are 3 images. Only ONE is AI-generated.
Your job is to tell us if it's the one on the Left, Middle, or Right.
Submit your guesses by Sat, Feb 28 at 11:59pm UTC. As usual we'll be revealing the answers on Sunday!
If you got 3+ right, we’ll share how to claim your reward next week.
@PerleLabs Middle The middle image (the Leaning Tower of Pisa) is most likely AI-generated. The lighting and texture look slightly too uniform and polished, and the structure appears overly clean and symmetrical despite the real tower’s subtle irregularities and surface wear.
Left is genuine. Reasons:🔸 The lighting looks subtly irregular and realistic, particularly on the balloon panels and the basket.
🔸 The balloons farther away appear smaller and slightly blurred, aligning with natural atmospheric depth.
It's the final week of Human CAPTCHA: Round 7
You made it this far… but we’re not going easy on you. One of these is a real photograph of a hot air balloon, the rest is AI-generated. Can you tell which one it is?
Respond with Left, Middle, or Right and tell us why below!
Submissions close Sat, Feb 28 at 11:59pm UTC. Answers revealed Sunday.
@PerleLabs Left is genuine.
Reasons:🔸 The lighting looks subtly irregular and realistic, particularly on the balloon panels and the basket. The balloons farther away appear smaller and slightly blurred, aligning with natural atmospheric depth.
Time to think INSIDE the box 🐾
Our new task is live: The Perle Labs Pet Bounding Box Quest
You’ll review pet images and draw a tight bounding box around the pet’s head.
These annotations train object detection models, so clean, consistent boxes matter.
- 20 tasks
- 100 points each
- 1,000 points to qualify
@PerleLabs The image on the left was generated by AI.The clouds appear excessively dramatic and somewhat artificial, while certain building details and reflections look slightly distorted and unrealistically polished.
Help train better AI – and earn while you do it.
I'm contributing on @PerleLabs, where verified human feedback powers real AI data.
Join me here: https://t.co/NaHT3K78yp
The left one is real.
The fur, lighting, and overall atmosphere look natural and slightly imperfect, like a typical wildlife photo. The right one feels too polished and cinematic — that kind of “too perfect” look is often a giveaway that it’s AI-generated.
Perle Human CAPTCHA #4
It's time to put your human judgement to the test again. One of these images of a tiger is a real photograph, one was generated by AI.
Which one is real? Reply "left" or "right" and tell us why ↓
This week's submissions close Sat, Feb 14th at 11:59pm UTC. Both answers will be revealed on Sunday 👀
@PerleLabs The left one is real.
The fur, lighting, and overall atmosphere look natural and slightly imperfect, like a typical wildlife photo. The right one feels too polished and cinematic — that kind of “too perfect” look is often a giveaway that it’s AI-generated.
Perle Human CAPTCHA #3 📷
One of these images of fruit is a real photograph, one was generated by AI.
Which one is real? Reply "left" or "right" and why 👇
You have until Sat, Feb 14th at 11:59pm UTC. This week's answers will be revealed on Sunday.
New Quest on Perle Labs: Logical Fallacy Detection 💬
It’s time to put your reasoning skills to the test. Here’s what you’ll do:
1) Read short debate-style arguments
2) Spot the flawed logic
3) Pick the fallacy type
Human-verified. Auditable. Sovereign. Start contributing: https://t.co/JQagfcGU7Z
Help train better AI – and earn while you do it.
I'm contributing on @PerleLabs, where verified human feedback powers real AI data.
Join me here: https://t.co/NaHT3K78yp
it is good project
Help train better AI – and earn while you do it.
I'm contributing on @PerleLabs, where verified human feedback powers real AI data.
Join me here: https://t.co/NaHT3K78yp