• Acquisition Underscores the Importance of AI Security in Modern IT Infrastructure

By Michael Novinson (MichaelNovinson)   

Cisco plans to purchase a startup led by a former Harvard computer science and mathematics professor to enhance the security of AI applications and infrastructure.

The San Jose, California-based networking giant said its proposed acquisition of San Francisco-based Robust Intelligence will ensure secure deployment of AI technologies by safeguarding AI applications against emerging risks and protecting AI models throughout their life cycles. The company’s automation and risk mitigation capabilities will help Cisco customers deploy AI while complying with industry standards.

“Our intent to acquire Robust Intelligence marks an exciting new chapter in our journey to secure the AI-driven enterprise,” said Cisco Security senior vice president and general manager Tom Gillis in a blog post. “Together, we’ll push the boundaries of AI security, empowering organizations to embrace AI’s full potential while safeguarding their critical assets.”

Robust Intelligence, founded in 2019, employs 79 people and has raised US$44 million in three rounds of outside funding, culminating in a US$30 million Series B round in December 2021 led by Tiger Global. The company has been led since its inception by Yaron Singer, a former Google and Microsoft research scientist who spent more than a decade as a computer science and mathematics professor at Harvard.

“‘AI risk’ is not a single problem with a single solution. It is diverse and complex, manifesting in every aspect of the AI adoption/development life cycle,” Robust Intelligence co-founder Kojin Oshiba said in a LinkedIn post. “With Cisco’s vast product portfolio in its security platform and its remarkable reach to pretty much every security team on the planet, we now have a very special key to unlock the next chapter to realize our mission.”

What Robust Intelligence Brings to the Table

The company’s platform mitigates risks for AI models such as data poisoning and prompt injection while ensuring that AI technologies meet industry standards including OWASP and MITRE ATLAS, according to Cisco. The acquisition of Robust Intelligence will help Cisco integrate advanced AI defense features into its existing security and networking products, giving customers better control and visibility over AI traffic.

Inserting Robust Intelligence into Cisco’s security and networking products will provide visibility into all of a customer’s AI traffic, helping them more easily build, deploy and secure AI applications. The technology mitigates risks such as prompt injection, data poisoning, jailbreaking and unintentional model outcomes while mapping detections and testing to industry and regulatory standards, according to Gillis.

Gillis said Robust Intelligence’s technology and expertise will accelerate existing Cisco bets in areas such as simplified configuration, attack prediction, adaptive policy enforcement and user experience. He said Robust Intelligence’s focus on automated model assessment and proactive AI risk mitigation will enable customers to secure their networks, AI applications and users all in one place.

This is Cisco’s first security acquisition in 2024 following an incredibly active 2023, highlighted by the US$28 billion purchase of SIEM stalwart Splunk. In addition to Splunk, Cisco purchased open-source networking and security provider Isovalent, predictive and generative AI security firm Armorblox, startup Lightspin to prioritize and remediate cloud issues, and Valtix to enforce policies across public cloud environments.

“Security leaders need purpose-built solutions they can trust to keep up with the new paradigm of AI risk,” Singer said in a blog post. “Our team is excited about the integral role our platform will play in shaping the future of the Cisco Security Cloud, providing end-to-end AI security and leading the AI revolution with Cisco.”