Verbit’s data policy: How we protect your data in the age of AI

Verbit is a verbal intelligence and voice AI platform trusted by more than 3,000 businesses and institutions. Our mission is to make all verbal information and experiences accessible, insightful, and useful. That mission depends on something just as important as accuracy: customers trusting us with both live events and recordings that often contain sensitive, confidential, or personal information.

Since our founding in 2017, Verbit has combined artificial intelligence (AI) with domain-trained expertise and human involvement to deliver highly accurate transcripts and captions. Verbit’s platform has evolved to support live and post-production captioning, transcription, translation, audio description, dubbing, and note taking. We pair our proprietary automatic speech recognition (ASR) engine, Captivate™, with a technology layer that dynamically adapts terminology to each customer’s domain and use case.

Alongside our proprietary ASR engine, Verbit integrates generative AI (Gen AI) and large language models (LLMs) into select solutions to enhance our transcription and analysis capabilities and help customers turn their content into actionable insights.

AI tools are now everywhere – many of them free and a click away. That convenience raises fair questions about accuracy, intellectual property, confidentiality, privacy, and “hallucinations.” Every AI provider should answer those questions, and it’s especially critical for providers like Verbit, whose offerings combine significant generative and non-generative AI technology.

This page explains how Verbit handles your data, how these two types of AI differ in our solutions, and why that difference matters when you decide to connect an AI-driven tool to your regular Zoom or Teams calls or upload sensitive recordings to a provider’s platform.

Since its founding, Verbit has incorporated ASR technology into its core offerings, often combining it with the power of human intelligence to provide captions and transcripts of the highest quality. ASR technology is an example of non-generative AI.

More specifically – and unlike Gen AI-based technology, which we’ll explain further below – ASR engines output predictions based on statistical properties derived from the input training data.

Another example of non-generative AI is image categorization. An image categorization engine could, for example, be trained to distinguish between images of birds and airplanes. Such AI engines can never produce any output that contains data identifiable to the material used to train them. They merely learn to distinguish between the types of objects (birds and airplanes in this context) from many different examples of each of these objects. They do this by calculating and aggregating properties whose values are different for birds and airplanes.

These engines provide a statistical prediction when presented with a new image of one of these objects, categorizing the image as a bird or an airplane. However, such non-generative AI engines would be unable to generate an example of a bird or an airplane. They certainly could not provide output comprising an image from the training set.

ASR technology is similar. The output of an ASR engine comprises the most likely sequence of words that corresponds to the input audio signal. This is a prediction based on statistical properties derived from huge amounts of input training data.

In addition to ASR, Verbit also utilizes predictive AI models that enrich transcription output with auxiliary information. These models include speaker change detection, acoustic event identification (such as laughter, music, or applause) and spoken language detection. Like ASR, these models are non-generative in nature. They operate by analyzing patterns and statistical features in the input audio, such as changes in pitch, energy or spectral composition, to make predictions about specific attributes of the audio. Importantly, these models do not reproduce or generate content based on training examples and, therefore, they do not present any risk of intellectual property leakage or the inadvertent exposure of sensitive customer information. Their sole purpose is to augment the accuracy and usability of ASR output, ensuring better structure and richer metadata without compromising privacy.

You can read this helpful piece to learn about the evolution of ASR technology as a whole.

Verbit trains its proprietary ASR engine and ASR enrichment models with input consisting of short audio segments (each are a few seconds in length) together with their corresponding transcription. These inputs are used to correlate statistically between the acoustic speech input signal and the most likely sequence of words or letters it represents.

Verbit’s training data for its ASR engine consists of both publicly available data sets and media submitted by customers which trains the performance of their content alone. Training is done with many thousands of hours of audio. The data first undergoes statistical feature extraction, such as acoustic and spectral properties of short segments of the speech signal, phonetic information used to distinguish between different sounds and duration of pauses. Those features are accumulated together, so the model can learn the probabilities of the different letters or word outputs that correspond to the acoustic input.

Customers benefit from the strong customization and adaptation techniques Verbit employs, together with the fine segmentation Verbit performs over linguistic content and acoustics. Verbit invests significant effort in analyzing recognition accuracy over different customer verticals and tiers. This ensures that we use the model, vocabulary and configuration best suited for each customer and session. This is achieved by analyzing customers’ industries, spotting possible gaps, retraining and adapting the models to specifically address any specific customer need. These include specialized terminology, unique acoustic conditions, and specific accents that are frequent in the customer’s data or detailed formatting requirements.

Offering customers the opportunity to upload previous recordings of similar events, key terms and spellings, and customer-specific or domain-related content enables Verbit to provide a more seamless and tailored transcription service that best fits our customers’ needs and use cases. In essence, the utilization of customer data not only refines the technical aspects of the ASR output for each customer but also fosters a more reliable and customer-specific transcription output.

When we refer to “generative AI,” we mean AI models that can output both original and creative content – for example text, images, music or computer code in response to input prompts.

We differentiate between two sources of data which Gen AI (including the Gen AI capabilities in Verbit’s solutions) works with:

1. The data that the model was trained on (Training Data)
2. The data that the model uses when asked to perform a specific task – for example the transcript of a lecture the model is asked to summarize (Input Data)

Gen AI output may reflect (but generally does not duplicate) some of the data it consumes from both sources.

For example, when prompting for a summary of a lecture, the Input Data is the full transcript of the lecture. The Gen AI model provides a short text with the main ideas expressed in the lecture transcripts, using wordings that appeared in the Input Data transcripts as well as rephrasing some of the content of the lecture.

The rephrasing may resemble summaries and texts that were part of the Training Data set used to train the model. This means that it is possible the Gen AI model may output excerpts of the Training Data, excerpts of the Input Data and/or (for the most part) original outputs not readily recognizable as either Training Data or Input Data. Because of the theoretical possibility that Training Data may later appear – potentially unpredictably – in outputs from the trained Gen AI model, there is a legitimate concern that intellectual property, confidential information or personal data found in the Training Data may be exposed later.

For this reason, Verbit’s strict policy is to not use customer data as Training Data for Gen AI models, except where the customer has specifically requested a custom fine-tuned model. Such custom fine-tuned models are made available only to that customer (whose data was used for the fine-tuning) and not to any other customers. Verbit further ensures that any third-party Gen AI services it uses comply with these restrictions.

Free transcription apps, AI notetakers, and consumer chatbots can turn a recording into a transcript or summary in seconds. What they don’t always make clear is what happens to your content afterward. When you upload a deposition, a board meeting, a student hearing, an HR interview, or an unreleased broadcast to a free tool, that content can leave your organization’s control.

Here’s what’s at stake:

• Your content may train someone else’s model. Some free and consumer AI services’ terms allow them to use uploaded content to improve their models, sometimes by default unless you opt out. As explained above, Training Data can later surface in Gen AI output.

• No contract, no accountability. Free tiers rarely come with a Data Processing Agreement, defined retention and deletion terms or independent security audits you can review.

• Shadow AI is expensive. In IBM’s 2025 Cost of a Data Breach Report, one in five organizations reported a breach involving shadow AI – AI tools used without IT approval or oversight. High levels of shadow AI added about $670,000 to the average breach cost, and 63% of organizations surveyed had no AI governance policies in place.

• Compliance exposure. Recordings often contain personal data, privileged attorney-client communications or student records. Sending them to an unvetted tool can put your organization at odds with privacy laws, institutional policies and professional obligations.

Before uploading any recording to a free captioning or transcription solution, ask: Will my content train AI models? Who can access it? Where is it stored, and for how long? Can I delete it? Has the provider been independently audited? Our guide to transcription data security includes a full checklist for evaluating vendors.

With Verbit, you have clear answers to every one of those questions.

For all AI solutions, Verbit implements the following privacy protections:

• No cross-referencing: Models cannot access or learn from other customers’ data

• Encryption: All data is encrypted both in transit and at rest

• Access controls: Strict access controls ensure only authorized personnel can access customer data

• Audit logging: All data access and processing is logged for security monitoring

These AI-specific protections sit within Verbit’s broader security and privacy program, which includes:

• ISO 27001:2022 certification and SOC 2 Type II attestation for applicable services (learn more in our SOC 2 compliance guide)

• HTTPS for data in transit and AES-256 encryption for data at rest

• Two-factor authentication, role-based permissions and least-privilege access for employees

• Annual independent penetration testing

• A Data Processing Agreement, Standard Contractual Clauses and participation in the EU-U.S. Data Privacy Framework (including the UK Extension and Swiss-U.S. framework) for cross-border transfers

• Session-level processing of voice-derived data, such as voiceprints, which is deleted when the session ends unless a customer opts into a feature that retains it

• The ability to request removal of your data at any time through your account manager

Whether we are providing non-generative AI, Gen AI, or any other solutions to our customers, our priority is the security and protection of their data. An overview of our security program and certifications may be found on our Trust Page, while our privacy practices are described in our Privacy Policy.

Further, considering the fundamental differences between various AI technologies, we have taken a comprehensive approach to how we process customers’ data in the context of each solution type. We’ve found that the real blocker to AI adoption isn’t trust, it’s control, and we’ve set up our platform and AI capabilities to work accordingly.

With respect to our use of customer data for training our ASR engine and as described above, any information that might be identifiable to an individual is stripped out of the trained ASR models. Because ASR technology is non-generative AI, the output of Verbit’s ASR models cannot contain any segments of the data that was used to train the models, so there is no risk that customer data, intellectual property, or confidential information will be exposed in the output. We may also use LLMs for the purpose of enhancing the accuracy of ASR engine output. Consistent with Verbit’s policy toward Gen AI, we do not use customer data for training LLMs in this context either.

With respect to our Gen AI capabilities and LLM-based services, we maintain strict policies to ensure customer data privacy:

• We use LLM infrastructure that is designed with multiple layers of security and logical separation to ensure each tenant’s data and interactions are isolated from others.

• No cross-customer data sharing or learning occurs

These measures maintain the highest standards of privacy and security across all our services. That’s the real difference between a free tool and a trusted partner. With Verbit, you know how your data is used, who can access it with user hierarchies in place, and how it’s protected – so your team can capture live events, understand the core items captured with incredible insights flagged and available, and act on the insights delivered with full peace of mind.

Verbit’s secure platform supports organizations with leading AI across industries, including legal teams using Legal Visor and Legal Capture, colleges and universities using Campus Complete, government agencies using Civic Complete, and media companies using Captivate Post Plus, among our other captioning, transcription, dubbing, audio description, and localization solutions.

For procurement and compliance reviews, find our security program on the Trust Page, our privacy practices in the Privacy Policy, our customer agreements on the Legal page, and our accessibility conformance reports on the VPAT page. Questions? Contact our team.