#TrumpTariffs #BinanceAlphaPoints #LaunchpadWars
OORT's AI image dataset has reached the front page of Kaggle in multiple categories, highlighting the growing demand for high-quality, community-sourced training data.
An AI training image dataset developed by OORT, an AI solutions provider, has achieved significant success on Google's Kaggle platform.
OORT's list of diverse Kaggle tools was published in early April, and since then, it has topped the list in multiple categories. Kaggle is a Google-owned online platform for data science and machine learning competitions, learning, and collaboration.
“Kaggle rankings on the front page are a strong social signal, indicating that the dataset is attracting the right communities of data scientists, machine learning engineers, and practitioners,” Ramkumar Subramaniam, a core contributor to the cryptocurrency AI project OpenLedger, told Cointelegraph.
Max Lee, founder and CEO of OORT, told Cointelegraph that the company has “observed promising engagement indicators that demonstrate the early demand and relevance” of the training data collected through the decentralized model. He added:
“The organic interest from the community, including active use and contributions, demonstrates how community-driven, decentralized data pipelines like OORT can achieve rapid distribution and sharing without relying on centralized intermediaries.”
He also told me that OORT plans to release multiple datasets in the coming months, including a dataset for in-car voice commands, one for smart home voice commands, and another for deepfake videos, all with the goal of improving AI-powered media verification.
Related: AI Agents Are Coming to DeFi – Wallets Are the Weakest Link
Front page in multiple categories
Cointelegraph independently verified that the dataset in question made it to the top of the Kaggle platform's Artificial General Intelligence, Retail & Shopping, Manufacturing, and Engineering categories earlier this month. At the time of publication, these positions were lost following a potentially unrelated dataset update on May 6 and another on May 14.
The OORT dataset is on the first page of Kaggle in the Engineering category. Source: Kaggle
While praising the achievement, Subramaniam told Cointelegraph that it is "not a definitive indicator of its real-world applicability or enterprise-level quality." He added that what makes the OORT dataset stand out "is not just the classification, but also the source and motivation behind the dataset." He explained:
“Unlike centralized vendors that may rely on opaque pipelines, a transparent, token-based system offers traceability, community regulation, and the potential for continuous improvement, assuming proper governance is in place.”
Lex Sokolin, a partner at AI venture capital firm Generative Ventures, said that while he doesn't think these results are difficult to replicate, "they demonstrate that crypto projects can use decentralized incentives to regulate economically valuable activities."
Related: Sweat Wallet Adds AI Assistant, Expands into Multi-Chain DeFi
High-quality AI training data: a rare commodity
Data published by the artificial intelligence research firm Epoch AI estimates that human-generated text AI training data will run out by 2028. The pressure is so high that investors are now brokering deals that grant rights to copyrighted material to AI companies.
For years, there have been reports about the increasing scarcity of AI training data and how this could limit growth in the field. While the use of synthetic data (generated by AI) is increasing successfully, human data is still largely considered the best alternative, as it is high-quality data that leads to better AI models.
When it comes to images specifically used for AI training, things get more complicated as artists deliberately sabotage training efforts. Nightshade aims to protect their images from unauthorized use in AI training and allows users to distort their images, significantly impairing the models' performance.
Model performance per number of poisoned images. Source: TowardsDataScience
"We are entering an era where high-quality image data will become increasingly scarce," Subramaniam said, acknowledging that this scarcity is worsening as image distortion becomes more common:
“With the emergence of technologies like image steganography and anti-poison watermarking for AI training, open source datasets face a dual challenge: quantity and trust.”
In this context, Subramaniam stated that incentivized, verifiable, and community-sourced datasets are "more valuable than ever." He added that such projects "can become not just alternatives, but essential pillars for the alignment and provenance of AI in the data economy."
