Data validation company Truthset, on a mission to improve the accuracy of consumer data, has announced Record-Level Transparency, enabling data providers to classify demographic files based on score quality. The aim is to avoid treating all data files as if they are equal. It is available to all 22 Truthset Data Collective members.
Harvard University will release an AI training data set based on around one million books (public domain books only, unlike the controversial and much smaller Books3 dataset). This is part of the college’s new Institutional Data Initiative, funded by both Microsoft and Open AI. The data set is intended to help companies that can’t afford to assemble similar data on their own.
Reflecting the continued growth of cloud data platforms, Databricks says it is raising $10 billion (with a B) in Series J funding. The funding, led by Thrive Capital, values Databricks at $62 billion. Databricks reports over 60% growth YoY in recent quarters, some of which surely came from customer data.
Changes in marketing channels and data access have made advertising measurement more important, and challenging, than ever. Taking advantage of growing interest in the topic, Bain Capital is selling Kantar Group’s audience measurement division, Kantar Media, to global investment firm H.I.G. Capital for “approximately $1 billion”. Kantar Media accounts for 4,500 of Kantar Group’s 25,000 employees.