By Don Gossen, co-founder and CEO of Nevermined
Recent reports have predicted that publicly available, relevant data necessary for training AI models is expected to run out by 2026. The global AI market is projected to reach $1.8 trillion by 2030 — this creates a necessary need for additional ways to find and utilize other datasets. Not only that, but finding a way to monetize these AI models and datasets opens up a whole new creator economy, for both big business and individuals alike.
Take everyone’s favorite large language model (LLM), ChatGPT. While it recently gained access to current web data, it is still trained on the same publicly available, publically scraped data that other LLMs are trained on, like Google’s Bard. In order to differentiate and continue to grow their massive knowledge base, each LLM will need to train on private data not available to every other AI model out there.
Why is this necessary? To properly use these models, they are in need of relevant, context specific datasets. Not only that, but these models, and the publishers behind them, need the proper infrastructure to deal with today’s issue of copyright infringement, data privacy, and source referencing – leading to a need of building with web3 technology. Here’s why AI needs web3:
Revenue Generation
Monetizing data and AI models gives businesses, developers, and publishers a new stream of revenue generation. As developers are creating AI models, they are able to sell these on marketplaces to compensate for the work they’ve already done, which oftentimes includes research, human resources, and development of these models that can then be used by others for their own innovative technologies. Additionally, for businesses, a new stream of revenue can be created by monetizing their APIs in order to help these AI models train on non-public datasets. For example, Reddit recently closed it’s API, meaning AIs like ChatGPT can’t train and learn from the information being shared on subreddits. But Reddit could, in theory, decide to sell subreddit APIs on a marketplace platform, benefitting both their business and also the models hoping to train on that information.
Consent As A Driver
While monetizing AI & data models is a new and enticing revenue model for developers and businesses, one important thing to keep in mind is that there also needs to be informed consent from these developers and data publishers to grant access for additional model training and usage of their data and models. Building AI technologies on the blockchain ensures that this consent is being publicly tracked and traced, as well as provide the proper attribution for any materials used.
Innovation through Data Sharing and Collaboration
The collaboration possibilities for data and AI are seemingly limitless, creating net-new ways for everyone to work towards a shared and common goal through AI innovation. This will invariably lead to the development of more robust, comprehensive, and accurate AI models.
Accessibility and Scaling of Work
To make AI models and data more accessible to a wider audience, this collaborative value capture structure helps to scale the work being done. In turn, this can help AI models deploy on a larger scale, reaching more users and a wider audience or industry verticals.
With such a clear need for AI and data monetization for better growth in the space, how do we do it?
Blockchain technology is critical when it comes to the utility of these assets. Smart Subscriptions, which are tokenized versions of outdated subscription models, bundle multiple assets of data and AI models together, and give the asset owner the ability to set access variables like price and duration, is the best way forward to address the current issues listed above.
Through this type of monetization model, companies and publishers are able to leverage the transparent nature of blockchain technology, and provide the foundations for a trust-based AI asset registry.
TLDR; leveraging web3 technology to build AI gives a new growth vertical to help projects build in a transparent and scalable way, which helps to address the concerns that consumers and builders alike have. Both web3 and AI are the technologies of the future. It only makes sense that we combine the two to create reliable superpowers that will set up the next generation of innovation.
About the author:
Don Gossen is the co-founder and CEO of Nevermined, a data & AI company. Before founding Nevermined, he co-founded Keyko and Ocean Protocol, a decentralized data exchange protocol. Don has deep history and experience in data, previously also holding leadership positions at NTT Data. He is passionate about driving adoption and responsible progression of decentralized technologies through web3, data, AI, analytics, and Machine Learning. Don is based in Lisbon, Portugal.
Nevermined is a company based in Zug, Switzerland, that is building an open source Data & AI stack for Precision, Trust & Monetization that uses blockchain for data provenance & fine-grained attribution. In 2022, Nevermined raised $3M in a Seed round. The company is founded by CEO Don Gossen, CTO Aitor Argomaniz & CIO Dimitri De Jonghe, who met while creating Ocean Protocol, the world’s first decentralized data marketplace. In 2019, they founded Keyko over a shared belief in the power of digital ecosystems. They helped a variety of organizations develop web3 ecosystems, worth a combined value of over $2 billion. More information can be found at https://nevermined.io/
The views and opinions expressed herein are the views and opinions of the author and do not necessarily reflect those of Nasdaq, Inc.