Human Native AI is creating a marketplace for AI training licensing deals, streamlining the process of buying and selling AI training data
We must train AI systems and complex language models on vast quantities of data to achieve accuracy. However, they shouldn’t train on data for which they lack the necessary rights.
Last week, the Atlantic and Vox’s licensing agreements with OpenAI demonstrated that both parties are eager to secure these AI-training content licensing agreements.
Human Native AI is a London-based startup that is developing a marketplace to facilitate transactions between the numerous companies that are developing LLM initiatives and those that are willing to license data to them.
Its goal is to help AI companies find data to train their models on, while also ensuring compensation and opt-in for the rights holders.
Rights holders can submit their content at no cost and establish connections with AI companies to secure revenue sharing or subscription agreements.
Furthermore, Human Native AI assists rights holders in preparing and pricing their content, as well as monitoring any copyright infringements. Human Native AI charges AI companies for its transaction and monitoring services and takes a portion of each deal.
James Smith, the CEO and co-founder, attributed the inspiration for human-native AI to his previous involvement with Google’s DeepMind initiative.
In addition, DeepMind encountered difficulties in training the system due to a lack of sufficient high-quality data. Then he observed that other AI companies encountered the same problem.
Smith stated, “It appears that we are currently in the Napster era of generative AI.” “Is it possible to transition to a more positive era?” Is it possible to streamline the content acquisition process? Is it possible to provide creators with a certain degree of control and compensation? My mind was preoccupied with the question, “Why is there no marketplace?”
He presented the concept to his friend Jack Galilee, an engineer at GRAIL, during a stroll through the park with their respective children, as he had done with numerous other potential startup ideas. However, Galilee, in contrast to previous eras, advised them to pursue it.
Established in April, the organization is currently in its beta phase of operations. Smith stated that the demand from both parties has been extremely encouraging, and they have already signed a number of partnerships that they will disclose in the near future.
LocalGlobe and Mercuri, two British micro-VCs, led a £2.8 million seed round that Human Native AI announced this week. According to Smith, the organization intends to allocate the funds towards workforce expansion.
“I’m the CEO of a two-month-old company and have been able to get meetings with CEOs of 160-year-old publishing companies,” Smith said. “That suggests to me there is a high demand on the publishing side. Equally, every conversation with a big AI company goes exactly the same way.”
Although it is still in its infancy, the infrastructure that human-native AI is developing appears to be a gap in the expanding AI industry. The major AI actors require a substantial amount of data to train on.
It appears that providing rights holders with a more convenient method of collaborating with them, while also granting them complete control over the use of their content, is a viable strategy that can satisfy both parties.
Smith stated, “Sony Music has recently sent letters to 700 AI companies, requesting that they cease and desist.” “That is the size of the market, as well as the potential customers who could be acquiring data.” The number of publishers and rights holders could be in the thousands, if not in the tens of thousands. We believe that is the rationale for the need for infrastructure.
I also believe that this could be even more advantageous for smaller AI systems that lack the resources to negotiate an agreement with Vox or The Atlantic but still require access to data for training purposes.
Smith expressed their aspiration for this outcome, noting that the larger AI players have been involved in all notable licensing agreements to date. He is optimistic that human-native AI will contribute to leveling the playing field.
According to Smith, licensing content presents a significant challenge due to the substantial initial costs and significant restrictions on the collaborators with whom one can collaborate. “How can we reduce the barriers to entry and increase the number of buyers for your content?” We find that to be quite thrilling.
The other intriguing aspect of this is the potential for future use of the data that human-native AI accumulates. Smith stated that in the future, they will be able to provide rights holders with greater clarity regarding the pricing of their content by utilizing the platform’s transaction data history.
Additionally, it is an advantageous moment for human-native AI to initiate operations.
According to Smith, the ethical sourcing of data by AI companies, as well as the provision of receipts to substantiate their claims, will become increasingly important as the European Union AI Act continues to evolve and prospective AI regulation in the United States is considered.
“We are optimistic about the future of AI and what it will do, but we have to make sure as an industry we are responsible and don’t decimate industries that have gotten us to this point,” Smith said. “That would not be good for human society. We need to make sure we find the correct ways to enable people to participate. We are AI optimists on the side of humans.”
