From Cointelegraph

May 14, 2025 8:18:00 AM:

Decentralized AI solution provider OORT’s training data set on Google’s Kaggle platform saw success, ranking on the first page in multiple categories. Core contributors believe it engages the right community. OORT plans to release more data sets, including in-car voice commands and deepfake videos for AI media verification.

OORT’s data set reached the first page in Kaggle’s Engineering category, but rankings fluctuated due to updates. Experts caution that rankings may not reflect real-world adoption. OORT’s token-incentivized system offers transparency and community curation. AI projects can use decentralized incentives to organize valuable activities.

AI research firm Epoch AI predicts a scarcity of human-generated text AI training data by 2028. Investors now mediate deals for copyrighted materials with AI companies. Human data is seen as higher quality for AI models. Artists are sabotaging training efforts with image poisoning to protect their work.

Experts warn of a scarcity of high-quality image data for AI training due to techniques like image cloaking and adversarial watermarking. Open-source data sets face challenges in quantity and trust. Verifiable, community-sourced data sets are becoming more valuable for AI alignment and provenance in the data economy.

Read more at Cointelegraph: Decentralized OORT AI data hits top ranks on Google Kaggle