Automated PII Removal for Secure AI Data with Veritone

Post Views: 8

Veritone Strengthens AI Data Security with Automated PII Removal Tool

As the demand for artificial intelligence (AI) deployments and applications continues to grow, the need to ensure that AI training data is properly sanitized of personally identifiable information (PII) and other sensitive data has become increasingly pressing.

Addressing the Challenge

To address this challenge, Veritone has deployed its Veritone Redact tool in conjunction with Veritone Data Refinery (VDR) to automatically remove PII and sensitive data from AI datasets before processing.

Ensuring Compliance and Security

This move is designed to help organizations meet stringent industry compliance and privacy standards, while also enabling a broader range of companies to innovate and compete in the AI space.

By ensuring that AI training data is clean and free of sensitive information from the outset, Veritone’s VDR helps to mitigate the risk of data breaches and intellectual property theft.

According to Ryan Steelberg, CEO of Veritone, “We are committed to helping data-driven organizations protect their valuable assets and ensure that their data is used in a clean and ethical manner.”

Steelberg noted that Veritone’s Redact tool, which is traditionally used by public sector customers, including law enforcement agencies, is being leveraged to safeguard PII before it is processed.

The Need for Automated PII Removal Tools

Veritone Redact is an application designed for public safety and law enforcement agencies that automates the process of redacting sensitive information from audio, video, and image-based evidence.

The tool reduces manual redaction processes while increasing accuracy and minimizing errors, helping agencies to meet critical deadlines.

The Growing Demand for Compliant AI-Ready Datasets

The need for automated PII removal tools like Veritone Redact is becoming increasingly urgent, as the volume of data flowing into AI systems continues to grow.

According to the Stanford HAI 2025 AI Index Report, training datasets are doubling every eight months, raising pressing concerns about the legal and ethical risks associated with AI deployments.

A recent audit of over 1,800 text datasets found that frequent miscategorization of licenses on widely used dataset hosting sites was a common problem, with over 70% of datasets lacking proper licensing information.

This highlights the need for organizations to prioritize the use of compliant and ethically sourced AI-ready datasets.

Growing Demand for Veritone’s VDR Tool

Veritone is seeing growing demand for its VDR tool from both content owners and hyperscalers, with the volume of data processed increasing by 3.5 times in the second half of 2025 compared to the first half of the year.

This trend is expected to continue, as organizations seek to ensure that their AI deployments are secure, compliant, and respectful of data owner rights.

Note that I’ve wrapped the content in HTML, following the provided rules and formatting guidelines. I’ve also used the exact tags and formatting specified in the rules.