MineOS aims to illuminate the AI ‘black box’ for enterprise

The announcement comes on the heels of major AI advances like ChatGPT and Gemini that have sparked rising demand for oversight, according to Mine co-founder and CEO Gal Ringel. “AI governance has been top of mind for the tech world since ChatGPT was released, and I think we’ve seen that chain reaction more widely across society with the major privacy inquiries into genAI tech and the EU passing the AI Act,” Ringel said in an interview with VentureBeat.

While AI systems like large language models (LLMs) can provide tremendous benefits, their inner workings are often invisible to users. This had led to growing concerns around potential harms related to security, ethics, privacy and more.

The AI governance gap

Broadly speaking, frameworks and solutions for enterprises to effectively govern AI development and usage have been slow to arrive.

“The problem in many cases with AI opacity is that organizations don’t know they’re being used at all, which adds an unacceptable level of uncertainty and risk that we aim to resolve,” Ringel told VentureBeat.

This governance gap has caught the attention of lawmakers, with regulations like the EU’s AI Act imposing new assessment requirements. But creating and implementing controls remains a challenge for many companies.

Lighting up the ‘black box’

MineOS uses a combination of system scans and analysis of email metadata to detect nearly all data systems and AI tools used within an organization, both third-party and internally developed.

This enables companies to map their AI landscapes and comply with emerging regulations.

“Our email discovery investigates the metadata of emails looking for things like invoices and sign up notifications to detect additional data systems,” explained Ringel. “If an employee is using ChatGPT or Grammarly or another AI-powered system, there’ll be that email trace of it, and we can confirm that without looking at the content of the emails themselves.”

Once AI systems are discovered, MineOS customers can analyze risks, create governance rules tailored to new laws, and run audits to demonstrate compliance. This end-to-end approach aims to make AI governance seamless.

The competitive landscape

MineOS is entering an increasingly crowded space, facing off against tech giants like IBM, Microsoft and Google who are rapidly building out AI governance capabilities.

However, the company believes its privacy-focused DNA gives it an advantage. “Data privacy is and will be a core tenet of AI governance,” Ringel said. “As a product and company that’s taken an innovative approach on data privacy, the experience we have in automating critical data governance exercises like mapping has given us an incredibly strong baseline to manage AI governance.”

With $42.5 million in funding from top investors like Google’s AI fund, Gradient Ventures, MineOS hopes its privacy-first approach will make it a leader in illuminating the AI “black box” for enterprises worldwide. But in the high stakes race to define the future of AI governance, nothing is guaranteed.