A New Startup Wants To Pay People For Their Personal Data. They Get To Name the Price

A new startup is betting that consumers will be willing to sell their personal data directly to artificial intelligence companies instead of allowing technology platforms and data brokers to profit from it for free.

Verb, a data exchange platform, unveiled a marketplace that enables users to monetize selected personal information by selling it directly to businesses, including AI developers looking for fresh training data. The company said users will be able to determine what information they share, choose potential buyers and set the price for their data.

AI companies face growing demand for high-quality datasets to train increasingly sophisticated models. While many AI firms have relied on publicly available web content, licensed datasets or information obtained from publishers, app developers and data brokers, Verb argues that consumers themselves should be compensated for the data they generate.

"We're creating a place where people can monetize something that for decades they've given away for free," Verb CEO and co-founder Cyrus Beschloss told Axios.

Under the new marketplace, users install Verb's tracking application on their smartphones and can choose which categories of activity they are willing to share. The platform can collect information such as shopping behavior and social media usage, while allowing users to disable categories they do not want tracked, including search activity. According to the company, users who share more categories of information may be able to charge higher prices for their datasets, NBC News reported.

Verb said it will suggest pricing estimates based on the type and amount of information users choose to provide, although individuals ultimately decide how much to charge. Users can also block specific companies from purchasing their data if they choose.

The company expects AI laboratories to become its largest customers because they require massive volumes of fresh data to train future generations of AI models, Beschloss said. As publicly available internet content becomes increasingly exhausted for model training, proprietary behavioral information could become more valuable to developers seeking unique datasets.

The launch also highlights the growing debate over privacy and ownership of personal information in the AI era.

Verb said it excludes certain highly sensitive categories from collection and analysis, including passwords, text messages and health information. Beschloss acknowledged that safeguarding user information will be critical to the platform's success, saying the marketplace was built with security risks in mind.

The concept of paying consumers directly for their personal information is not entirely new, although previous efforts have struggled to gain widespread adoption. Companies already spend billions of dollars each year acquiring consumer information through advertisers, apps and data brokers, but individuals who generate that data rarely receive compensation.

Verb enters a market where several companies are experimenting with alternative approaches to AI data sourcing. Startup Vana, which emerged from research at MIT, allows users to pool their personal data and collectively decide which organizations can use it for AI training, while recruiting platform Mercor pays experts to generate and evaluate AI training data. Meta has also begun offering lower prices for some AI coding tools in exchange for permission to use customers' work to train its models.

The broader AI industry has also seen increased investment in marketplaces built around ethically sourced training data. Earlier this year, enterprise AI company Veritone introduced a marketplace designed to connect rights holders with AI developers seeking licensed datasets, reflecting the industry's growing interest in alternatives to scraped internet content.