WitnessAI is building guardrails for generative AI models

WitnessAI is building guardrails for generative AI models

WitnessAI is building guardrails by intercepting interactions between employees and custom generative AI models and enforcing risk-mitigating policies. These include restricting unauthorized AI use, redacting sensitive information, and protecting against prompt injections.

Demand for Controls

Businesses are aware of generative AI’s productivity potential but are also wary of its risks. Only 9% of companies feel ready to handle related risks, according to a Riskonnect poll, despite IBM claiming that 51% of CEOs are hiring for generative AI roles.

WitnessAI’s Method

WitnessAI is building guardrails that ensure AI tools are used safely and responsibly within enterprises. By intercepting communication between staff members and unique generative AI models, WitnessAI’s platform applies risk-reduction measures and security measures. This entails guarding against the improper use of AI techniques and removing private information from prompts. Additionally, the platform protects models from attacks that can lead them to create damaging content.

Security and Privacy Measures

While acknowledging some privacy issues, WitnessAI uses an encrypted, segregated platform to reduce them. To ensure data privacy and regulatory compliance, each customer has an own instance that is encrypted with their own keys.

Concerns about Employee Monitoring

As WitnessAI is building guardrails, it focuses on creating a secure and compliant environment for generative AI environment, balancing oversight with employee privacy concerns. Despite concerns about employee morale, it has 25 early corporate users and plans to go public in Q3.

Funding and Growth

By the end of the year, WitnessAI hopes to grow its workforce from 18 to 40 members after raising $27.5 million from GV and Ballistic Ventures. With competitors like AWS, Google, Salesforce, and startups like CalypsoAI, Caccia emphasises the strategic expansion plans and strong sales pipeline.

Addressing AI Adoption Barriers

The platform for AI seeks to remove obstacles to AI adoption by improving visibility, control, and security. To prepare for early deployments in June 2024, the company’s Secure AI Enablement Platform tackles AI privacy, governance, and security.


WitnessAI is building guardrails for generative AI models, ensuring ethical and responsible deployment of these powerful technologies. By establishing frameworks for accountability and oversight, WitnessAI empowers users to harness the potential of AI while mitigating risks, ultimately fostering a safer and more trustworthy AI ecosystem

Leave a Comment

Your email address will not be published. Required fields are marked *