Threat ecosystem of Generative AI
When we started dabbling with GenAI applications, we noticed that while GenAI was making a lot of business functions easier and faster, it opened a new surface area for threats and attacks. Threats which didn’t exist before the advent of GenAI, for example:
Prompt injection
Toxic content generated through GenAI systems
Abuse of GenAI systems causing financial loss
Training data leakage
System prompt leakage
Data exfiltration
and the list goes on.
We soon realised, these new attack vectors couldn’t be tackled by traditional cybersecurity measures.
Ruchir and I both come from a background of threat and risk management and we are firm believers that within enterprises, managing risk can catapult growth and profitability. We similarly believe that managing the risks associated with GenAI can help organisations harness the full potential of this powerful technology.
Announcing Seed Round for SydeLabs
Today, we proudly announce our seed round and are even more thrilled to work towards our mission of empowering enterprises to adopt GenAI confidently. We take immense inspiration from traditional leaders in the cybersecurity space such as Palo Alto networks, Cloudflare, Google and Microsoft.
We have raised a $2.5M seed round led by Galina Chifa from RTP Global. Our seed round also saw significant participation from Picus Capital led by Naman Jhawar. We are also excited to have some of the prominent names in AI and Security join this round - Kunal Shah, Mike Wiacek, Sai Srinivas, Shubh Malhotra, Davit Baghdasaryan and others.
“We are super excited about the potential SydeLabs has in setting new standards for AI applications that are as secure as they are ground breaking. We love the vision of Ruchir and Ankita and are happy to back them so early in their journey. SydeLabs' approach to AI security showcases the kind of forward-thinking application of technology we champion at RTP Global” - Galina Chifina, Partner, RTP”
What is SydeLabs?
SydeLabs is a comprehensive AI Security and Risk Management platform. We are building
SydeBox : an Automated Red Teaming solution for GenAI systems
SydeGuard: an intent based AI firewall for run-time protection of GenAI systems
SydeComply: our compliance gap assessment and enablement solution
All in all, one unified platform that can be leveraged for mitigating all risks associated with your AI systems.
The Present: What SydeLabs does to secure GenAI?
We went live with launch of beta version of SydeBox on 1st March, 2024. Since then, SydeBox has detected over 10K vulnerabilities in over 50 AI system across 15+ enterprise design partners. We are soon launching SydeGuard in beta and onboarding new adopters everyday.
We have built a world class threat intelligent database which is updating every single day to cover for new attack vectors and techniques unfolding.
We have also run our Automated Red Teaming solution on multiple open source, foundational and fine tuned models and have detected vulnerabilities across the board. We would soon be publishing a detailed comparison of models on safety and security alignment.
What is in SydeLabs' roadmap for GenAI's future?
While building SydeLabs we’ve realised that in this space you continuously need to stay ahead of the bad actors, always pre-empting attack modus operandi and techniques.
With an ever evolving landscape - we’ll be allocating our resources to building the best threat detection and prevention modules which increase the defense of AI systems every single day.
Our early partners are seeing immense value in detecting vulnerabilities in their AI systems even before going live. With the launch of SydeGuard we would further augment the line of defense for AI systems and our partners.
We are also exploring partnerships with other players in the GenAI ecosystem and see meaningful synergies in co-bundling solutions for GenAI.
We’ve had immense support from our early supporters, advisers and the community alike and we believe we can give back to the community by building solutions which boost GenAI adoption.
“We also need to protect our systems from blatant attacks, instead of only focussing on subtle roleplaying or jailbreak techniques. Box and Guard are both important products to have at any stage of our AI development.” — CISO at a leading Ecommerce brand
We are grateful to our early adopters who’ve placed their trust in us and the solutions we are building. This gives us the fuel to continually improve and build solutions which indeed overcome any hesitation around adopting and deploying GenAI.
How to Access SydeLabs platform
Request access to our beta program and try your hands on SydeBox to detect vulnerabilities in your AI. You could also contact us at contact@sydelabs.ai to discuss further on our plans and alignment with your GenAI strategy.
