GV, Ballistic Ventures Lead $27 Million Round in WitnessAI

<p><strong><span class&equals;"legendSpanClass"><span class&equals;"xn-location">SAN MATEO<&sol;span><&sol;span><&sol;strong> &&num;8212&semi; WitnessAI&comma; creator of the first enablement platform for safe AI use&comma; has secured a <span class&equals;"xn-money">&dollar;27&period;5 million<&sol;span> Series A round&comma; co-led by GV and Ballistic Ventures&period; The company had previously been incubated by Ballistic Ventures&comma; beginning in 2023&period;<&sol;p>&NewLine;<p>Generative AI and large language models &lpar;LLMs&rpar; are seen as a transformative technology for enabling innovation across every business function&comma; yet few enterprises have actually adopted it fully&period; <u>A recent survey<&sol;u> showed that&comma; while 75&percnt; of enterprises have experimented with generative AI&comma; only 9&percnt; have actually deployed it widely&period; Privacy and governance issues were among the top barriers cited&period; Common AI governance issues include&colon;<&sol;p>&NewLine;<ul >&NewLine;<li><b>Lack of visibility<&sol;b> &&num;8211&semi; With new AI chatbots and projects appearing on the web each day&comma; IT organizations are often unable to track which AI systems their employees are accessing&comma; and what employees are doing with those systems&period; An internal WitnessAI survey indicated that nearly 90&percnt; of CISOs struggle to get a clear picture of employee AI usage&period;<br class&equals;"dnr" &sol;><br class&equals;"dnr" &sol;><&sol;li>&NewLine;<li><b>Lack of control <&sol;b>&&num;8211&semi; AI presents a new set of privacy and compliance challenges&comma; such as preventing LLM training data from one client from being used to serve a different client&comma; ensuring that employees can&&num;8217&semi;t illegally access customer data within a private LLM&comma; or blocking company IP from being shared with a public LLM such as ChatGPT&period;<br class&equals;"dnr" &sol;><br class&equals;"dnr" &sol;><&sol;li>&NewLine;<li><b>Lack of protection <&sol;b>&&num;8211&semi; LLMs create a new attack surface&comma; putting enterprises at risk of data or financial loss&period; Prompt injection&comma; jailbreaking and hallucination are a few of the common security risks from LLMs in use today&period;<&sol;li>&NewLine;<&sol;ul>&NewLine;<p>WitnessAI addresses these challenges to enable safe&comma; effective use of AI within the enterprise&period; The WitnessAI Secure AI Enablement Platform includes modules for observability of AI use&comma; policy enforcement and governance of AI use&comma; and protection of employees&comma; customers and enterprise data from AI use&period; The company will begin early deployments in <span class&equals;"xn-chron">June 2024<&sol;span> and is already working with more than 20 design partners across multiple industries&period;<&sol;p>&NewLine;<p>&&num;8220&semi;Building a safe AI enablement platform that works at enterprise scale requires a purpose-built team with expertise in AI and ML&comma; red and blue teaming&comma; and web-scale platform deployment&comma;&&num;8221&semi; said <span class&equals;"xn-person">Rick Caccia<&sol;span>&comma; WitnessAI CEO&period; &&num;8220&semi;We&&num;8217&semi;ve built something that works in the most demanding environments&comma; including global organizations and across clouds&period; The response from CISOs we&&num;8217&semi;ve spoken with has been tremendous&period;&&num;8221&semi;<&sol;p>&NewLine;<p>WitnessAI is deployed as an isolated&comma; cloud-based instance for each customer&comma; encrypted with their own keys to ensure data privacy and regulatory separation&period; The platform intercepts activity between users and LLMs&comma; applies control policies&comma; and ensures data protection and user safety – all with millisecond latency and requiring nothing to be installed on the users&&num;8217&semi; devices&period; Achieving this requires an integrated design that blends new techniques in machine learning&comma; offensive and defense security for AI activity&comma; and enterprise-scale web services deployment&period;<&sol;p>&NewLine;<p>&&num;8220&semi;For most enterprises&comma; the AI journey is just beginning&comma;&&num;8221&semi; said <span class&equals;"xn-person">Karim Faris<&sol;span>&comma; General Partner at GV&period; &&num;8220&semi;The need for guardrails around AI use&comma; including visibility and control of that use&comma; will only grow&comma; and GV is excited to partner with the WitnessAI team as they make AI safe for the enterprise&period;&&num;8221&semi;<&sol;p>&NewLine;