Artificial intelligence has an insatiable hunger for high-quality data, and there are real opportunities for anyone who's able to provide it in sufficient volumes. Companies like Alexandr Wang's former startup, Scale AI, have already made multi-billion-dollar businesses out of labeling and curating fresh AI datasets, and now that same opportunity is being extended to anyone who cares to join in.
Imagine spending a day strolling around the neighborhood, snapping photos of different colors and breeds of stray cats and dogs, carefully approaching them from different angles to get the full perspective. Or simply strapping a GoPro to your head while pottering about in the kitchen, cooking meals, washing the dishes, and cleaning up afterwards. These are just some of the ways people can go about collecting the desperately needed data for training new AI models, but they're not doing it for free.
A startup called OORT is in the business of making data collection a nice little sideline for anyone with a smartphone and some spare time on their hands. It likes to call itself the "TikTok for data," and with the launch of its new DataHub Launchpad, it's giving everyone the chance to make a little pocket money.
It's essentially a kind of crowdsourcing hub for data collection. AI developers or researchers can post "data tasks" on its platform and specify the kind of data they need. For instance, if they're building a new vision model, they may very well want photos of cats and dogs and many other kinds of animals to train it on. So they outline what they need, and then anyone who's inclined to do so can grab their smartphone, head off outdoors, and take as many good-quality photos as they can, upload them to OORT's DataHub Launchpad, and get paid in cryptocurrency.
DataHub Launchpad also extends to social media campaigns and provides a way for influencers and companies to launch Data Tasks that are similar to "quests" found on platforms such as Zealy and Galxe. For instance, an influencer might request their followers to post images of screwdrivers to build up an image dataset, rewarding followers for each one they post. Once the dataset is complete and verified, OORT will help those influencers to monetize it by listing them on major data marketplaces, generating new revenue streams for them.
OORT believes this kind of data collection could really catch on at a time when AI models are becoming much more complex. While earlier models, such as GPT-3 were trained on fairly random data, with OpenAI scraping the entire internet to get its hands on as much content as it could, that won't do for today's sophisticated models. The most advanced reasoning models are trained in complex reinforcement learning environments that require strategic data collection. The more complex those environments are, the more precise the data requirements become. In many cases, they can only be fulfilled by people who set out specifically to collect that data.
Read more: "This Decentralized AI Data Provider Aims to Solve the Centralized AI Black Box Dilemma," by Tech Times
That's where the incentives come in. There are billions of people in the world, and more than a few of them are unemployed, or else looking for a way to earn a little extra cash. By tapping into a global pool of casual labor, OORT thinks it can help to fix a critical shortage of AI training data that has got many people very worried.
OORT makes sure the data is good. After the contributors collect it and upload it, others will be able to chip in by verifying and labeling that data, ensuring that it meets the specifications of the data task in question. When the AI project in question is satisfied, they'll release the crypto they deposited in an escrowed smart contract, so the contributors can get paid for their work.
The DataHub Launchpad supports different kinds of data collection. Besides typical AI training datasets and social media tasks, it also offers survey campaigns, where users can sit and answer questions to provide companies with the insights they desire to enhance their product development, advertising techniques, or whatever else they need.
OORT's founder and Chief Executive, Dr. Max Li, said the idea of democratized data collection is a powerful one that can transform the AI industry, solving not just shortages, but also problems such as model bias. By incentivizing everyone to take part, the data generated by its community will be extremely diverse, representative of people from every part of the world.
"Projects get the diverse, human-generated data they desperately need, while users get rewarded for their time and contribution," Li said. "It's a win-win for everyone, and we're excited to see a vibrant community grow around this concept."
ⓒ 2025 TECHTIMES.com All rights reserved. Do not reproduce without permission.
