AI vendors are relearning the oldest rule in the data business (www.techradar.com)

AI vendors are grappling with a fundamental issue in data management: how to fairly handle data pooling for model training while ensuring customers understand the implications. Drawing parallels to historical data cooperatives, software companies are finding that, unlike traditional systems where data usage was straightforward, AI necessitates a more complex understanding of data sharing. For instance, HubSpot faced backlash when it attempted to implement a default opt-in for sharing customer enrichment data across accounts, leading to concerns about unconsenting contributions benefiting competitors without any clear return for the data providers. This situation highlights the need for transparent data governance in the AI/ML sector. As companies increasingly incorporate AI features, they inadvertently shift towards becoming data companies, often without establishing clear terms on data use among clients. The challenge now lies in rewriting contracts to address data ownership, the collective enrichment from shared models, and what happens to a customer’s contribution if they opt out. As procurement teams push for clarity on these issues, vendors that proactively establish and communicate clear data economics will stand to gain a competitive advantage while enhancing trust with their clients.