BetterData taps blockchain to help create better synthetic data • TechCrunch
As the global The privacy regulatory landscape is becoming more convoluted and restrictive, engineering teams looking to use structured data to improve their products and AI models are being pushed to jump through many hoops to stay compliant.
Launching on stage at the TechCrunch Disrupt SF Battlefield startup competition, BetterData aims to help customers quickly generate representative, synthetic structured data so that technical teams can work with data in a compliant manner without waiting months for permission to use actual user data or generate their own.
The company’s product helps generate the AI data in a secure way that allows clients to upload real user data and securely transfer and convert it without a copy of the data landing on BetterData’s servers. User data is tokenized and stored on a blockchain that is only accessible with a user’s private encryption keys.
Image credit: Better Data
The generative data copy maintains the key properties of the original structured data while anonymizing and distorting the information. This enables teams to train models and create products capable of analyzing organic user data, but helps them avoid lengthy bureaucratic processes often required to access user data.
The start-up’s co-founders, managing director Uzair Javaid and CTO Kevin Yee, have backgrounds in AI data generation and blockchain security. They met at the Entrepreneur First program in Singapore.
The duo has raised $770,000 in funding and grants so far and is closing in on a seed raise.
“We’ve talked to hundreds of data teams … and they’re all facing the problem of access to data,” Yee told TechCrunch in an interview. “It takes a long time to access data under data protection rules … They are trying to innovate, but it takes so much time.”
Image credit: TechCrunch
The company announced on stage that it will expand the private beta after a series of successful pilot programs. BetterData specifically targets clients in the Banking Financial Services and Insurance (BFSI) world, as well as data and AI teams at technology companies.
Yeah and Javaid hopes that their product can not only help these teams stay compliant with the growing proliferation of privacy regulations, but can also help them avoid data attacks and leaks by tapping into encryption and the blockchain. The blockchain element will also allow customers to have an immutable access log and a complete record of data lineage so they can ensure that data is never tampered with.
For now, the company’s product focuses exclusively on processing and generating structured data, but as they build out the functionality, they plan to start generating text data using natural language processing models. They plan to launch a public beta of their cloud service solution by the end of this year.
