Why we are building Nevermined
Nevermined champions open Data and AI to lower the barrier of entry for AI adoption and advancement. This is running counter to the centralization of AI development and the control of training data.
Two critical concepts that will define AI adoption are:
- Data Availability
- AI Precision & Trust
To develop the big AI Models, a company needs plenty of data. But developing such AIs is not cheap, not fast, not democratic and definitely not sustainable.
That explains the trend towards more precision-focused techniques, like Transfer Learning or Federated Machine Learning, which can create hyper-relevant improvements, based on small data samples.
Specialized Precision AI models are incredibly valuable as they allow organizations to solve specific problems and optimize processes within their industry. The result is significant cost savings and improved efficiency based on accurate outputs that organizations can trust.
However, the need to boost speed, cost efficiency, accuracy and model reliability will need to be balanced with data access, privacy, data movement and sustainability.
So, funnily enough, the bottleneck is data. The precision of these AI techniques will be defined by the access it has to relevant data.
That relevant data typically doesn’t sit within one organization, but with partners, competitors, global teams or even clients. So, the arms race for walled data will be fierce, involving trust and incentives.
We believe that Web3 tech is the key enabler to unlock this walled data and to create an open ecosystem approach to how AI and Data develop and interact.