Connect with us

Tech

Licensed AI data marketplace startup raises £2.8m

Published

on

Human Native AI, a London startup offering what it describes as legal and ethical data for tech firms to train large language models, has raised £2.8m.

Founded in 2024, Human Native AI is aiming to solve the growing concern that many AI models unfairly use copyrighted data in the training phase.

As a nascent technology, the legality of using copyrighted art, music, photographs, writing and more to train generative AI models remains a hotly contested point.

Some firms, such as London’s Stability AI, have argued generative AI outputs are transformative and therefore using copyrighted data falls under existing fair use laws.

Others, including a former VP at Stability, claim AI models using data without the owners’ consent are effectively stealing and profiting off of other people’s work.

Human Native AI hopes to circumvent this problem with a marketplace of licensed content that developers can purchase to train their models.

The marketplace offers several options for licensing data, including revenue-sharing agreements, subscriptions and per-use payments.

The seed funding round for the startup was led by LocalGlobe and Mercuri.

“Rights holders are demanding greater control over how their works are being used to train AI systems,” said Alan Hudson, founding general partner at Mercuri.

“We think Human Native AI is the answer to this problem, preserving human creativity in the face of rapid technological developments.”

LocalGlobe partner Ziv Reichert added: “Individual, case-by-case licensing agreements are not practical, and the scraping and subsequent cleaning of data is not sustainable.

“Human Native AI’s marketplace provides a rich data environment for AI developers while protecting intellectual property rights, and ensuring fair competition and participation in the AI industry.”

Continue Reading