Data Mask

A privacy firewall for LLMs. Anonymise sensitive data when accessing cloud-based LLMs

A Generative AI chatbot built on your data in a snap

What is Data Mask?

Data Mask is a masking tool that keeps personal and corporate data private when interacting with cloud-based Large Language Models like GPT. It obscures sensitive information within data queries through customisable filters and alert system integration, enabling the secure use of AI without disrupting business operations.

What type of data does Data Mask conceal?

Personal data that often needs encryption for security includes:

- Name
- Address
- Email
- Telephone number
- Date of birth
- Passport number
- Driver's license number
- Credit or debit card numbers
- Social Security number
- Gender and sexual orientation

Tovie AI is secure AI

Only 9% of companies are prepared to manage generative AI risks, indicating a significant readiness gap. The technology is moving so fast that companies don’t know where to start

65% of companies cite data privacy and cyber security as their primary concerns regarding AI.

Risconnect, The New Generation of Risk Report 2023

Getting a handle on the risks of GenAI now is paramount to successfully embracing the technology as a value driver

Businesses need to reduce the chance of data breaches while using neural networks and external LLMs. The solution is implementing a data analysis gateway that substitutes sensitive information with fictional, yet semantically consistent, data.

Data Mask replaces sensitive data with placeholders encrypted for security during processing, restoring the original data after AI processing is complete. It protects various personal details, including names, addresses, and financial information. In software development, it also safely replaces IP addresses, tokens, and passwords.

Data Mask benefits

Enables utilisation of cutting-edge generative models, including GPT-4 Turbo, Google’s PaLM, Anthropic's Claude, Meta’s LLaMA, Vicuna and Falcon

Reduces costs by removing the necessity for hosting and managing private LLMs

Adheres to organisational security protocols and conceals sensitive information while preserving data context

How Data Mask works

Provides customisable data filtering policies to define rules for detecting sensitive information and the response when it's found

Uses machine learning-based Named Entity Recognition (NER) to identify sensitive data alongside rule-based methods

Conducts data masking by substituting confidential elements in a query with placeholders, ensuring the request's meaning remains intact

Records logs of queries to and responses from external systems and can be integrated with a company's alerting systems

Data Mask features

Compatible with the Open AI API protocol for seamless operations

Boasts a versatile architecture that supports component deployment both in the cloud and on-premises

Customisable for identifying and obscuring specific data elements such as names, passport details, credit card numbers, and email addresses

Integrates effortlessly without the need for altering standard business procedures or complex additional integrations

Elevate your business with generative AI—start your journey to success with our expert guidance!

FAQs

What is data masking software?

Data masking in generative AI is a technique that replaces sensitive information with fictional yet relevant placeholders, allowing the safe use of AI models like Large Language Models without compromising privacy. It utilises customisable filters and machine learning identification and integrates with alert systems to preserve data integrity and maintain security compliance.

How does data masking work?

In generative AI, data masking tools identify and anonymise sensitive data before processing, allowing the AI to generate outputs without exposing real data. This is achieved through rules or algorithms that detect and replace confidential elements with synthetic data. So basically, this is a data leak prevention software.

What is data masking in GDPR?

Data masking creates a convincing yet artificial version of company data. It aims to secure sensitive information by providing an operational stand-in for scenarios that do not require real data, like training sessions, sales presentations, or software trials.

Data Mask assists with GDPR adherence by obscuring personal details, sticking to minimal data use, and guaranteeing safe data handling. It helps establish data protection and facilitates lawful cross-border data transfers while maintaining individuals' data rights. However, it should be part of a broader GDPR compliance strategy, not the sole solution.

What type of data does Data Mask conceal?

Personal data that often needs encryption for security includes:

Name
Address
Email
Telephone number
Date of birth
Passport number
Driver's license number
Credit or debit card numbers
Social Security number
Gender and sexual orientation

Which entities can be replaced in the software code?

In software coding, Data Mask can also secure IP addresses, tokens, and passwords by replacing them with alternative secure values.

Can the rules and entities be configured manually?

Yes, any rules and entities can be configured manually. Please get in touch with our team at contact@tovie.ai to learn more.

Can Data Mask provide different access levels for different employees?

Yes, Data Mask permits setting up flexible filtering rules for sensitive information. This means different access rights can be configured for employees, allowing certain individuals to work with personal data fully, partially, or not at all.

Do you have an API that can be used to address internal tasks - such as masking logs or database replicas?

Yes, Data Mask has that functionality. Don't hesitate to get in touch with our team at contact@tovie.ai to learn more.

Does Data Mask always mask data with the same placeholders or does it use random ones?

The data is changed to a placeholder according to the type of personal data. For example, the name is changed to PERSON_NAME_1, then demasked to its original form upon receiving a response from the LLM.

Who uses Data Mask?

Data masking solutions are primarily used by the following groups:

Businesses and Enterprises: To protect customer and employee data while performing data analysis, application testing, and AI activities.
Healthcare Organizations: For securing patient health information while still allowing for research and analytical uses.
Financial Institutions: To ensure customer financial information is kept confidential during processing and analytics.
Government Agencies: For protecting sensitive citizen data while maintaining government operations that require data handling.
Technology Companies: Especially those developing or using AI and machine learning to train models without exposing sensitive information.
Legal and Compliance Teams: To manage and enforce data privacy regulations and company data governance policies.
Researchers and Academic Institutions: Conducting studies with private data that must remain confidential to respect participant privacy.
IT and Security Teams: To implement and oversee data protection strategies across their organisations’ information systems.