Nightfall AI introduces “Firewall for AI” to enhance security of generated AI applications

Nightfall AI, a cloud data protection startup, today announced the release of Firewall for AI, a new product designed to protect generative artificial intelligence-based applications and data pipelines.

The service aims to address the issues of sensitive data leakage and instant injection, two of the biggest risks faced by companies that host publicly available large-scale language models or leverage publicly available generative AI services from companies like OpenAI and Google LLC. Nightfall's firewall for AI aims to address these concerns through comprehensive security, operational and content guardrails for AI models and applications.

Nightfall's Firewall for AI acts as a client wrapper that protects enterprise and customer interactions with generative AI-based applications and data pipelines. The service scans automated workflows and data pipelines to remove sensitive personally identifiable information, payment card information, personal health information, and sensitive information to prevent sensitive data from being exposed to his LLM. Prevent. This also ensures compliance with standards such as the European Union's General Data Protection Regulation and the California Consumer Privacy Act.

A firewall for AI also protects against malicious attacks such as prompt injection. This is done by detecting malicious content and ensuring proper language usage, code usage, response relevance, and sentiment analysis. Firewall for AI further maintains data quality by extracting proprietary, malicious, harmful, and irrelevant content from datasets.

“Traditional security solutions are often inadequate because they rely on simple regular expressions, heuristics, or outdated open source models,” said Rohan Sathe, Chief Technology Officer at Nightfall AI. I am. “Our Firewall for AI leverages an AI-native, enterprise-grade detection engine and fine-tuned DLP models to deliver unparalleled accuracy, throughput, and response time.”

Nightfall's firewall for AI seamlessly integrates into existing workflows through an application programming interface and software development kit, enabling customers to continuously monitor AI interactions and automatically detect and mitigate potential risks in real time.

The company is a venture capital-backed startup that last raised $40 million in a Series B round in August 2022. Investors include Bain Capital Ventures Partners LLC, Venrock Associates LP, Pear Venture Partners LLC, and several private investors, including the “Ant-Man” actor. Paul Rudd.

Photo: Nightfall AI