RSA Innovators Dinner "AI x CyberSecurity" Recap

“Oversubscribed” is the word rarely heard recently in Silicon Valley tech industry. But it just happened in a historic Hillsborough California mansion, during the RSA Conference, the world’s largest cybersecurity conference.?

Immediately after Anchain.ai won the RSA Innovation Sandbox Finalists, I planned to host a small party to celebrate with my cybersecurity friends that’ve been supporting our mission. It’s an incredible honor to be the first Web3 security company on the RSA stage.?

So I started a small 10 person RSVP party list but soon it went out of control. “Are you invited to Victor’s party too?”

Eventually we oversubscribed, settling at 35 guests, the maximum capacity our private chef can handle, while keeping the quality of the conversations and culinary service top notch.?

As promised, I am publishing the summary of the AI x Cybersecurity dinner note, in the hope that our thought provoking conversations could inspire more!

What is a Jeffersonian style dinner ??

Jeffersonian Dinner tradition began at the home of Thomas Jefferson in the late 1800’s. Jefferson hosted dinners with thought leaders from various fields to converse and debate on a topic, unlocking collective wisdom. It’s my favorite Silicon Valley event form factor !?

Topics and Demographics

The Jeffersonian style dinner was scheduled into 4 topics, led by the co-hosts:?

1. What is counterintuitive in AI x Cybersecurity?? By Victor Fang, Ph.D, CEO, AnChain.AI (RSA Innovation Sandbox), fmr Mandiant.?
2. Rise of Generative AI and its Impact on Cybersecurity. By? Richard Seiersen, CISO of? Resilience, Author of best sellers “How to Measure Anything in Cybersecurity Risks”
3. AI safety and responsibility. By?David Cass, CISO of GSR, fmr Federal Reserve Bank.?
4. AI x Cyber Investment areas. By? Chris McCann, GP of Race Capital, fmr Greylock.?

We have the honor to host these AI x Cybersecurity top minds, with the right mixture of 12% Innovation sandbox finalists, 27% researchers from Berkeley, Stanford, OpenAI, and 41% VC looking to invest in the Cybersecurity x AI sector.

Topic 1: What is Counterintuitive in AI x Cybersecurity.??

Every guest was required to prepare for this thought provoking question as an ice-breaker.?

• Low sophisticated attacks are stronger.? Hackers don’t have time to use GPT4 for social engineering, because there are plenty of low hanging fruits already.?
• Most security postures & teams are more immature than you think, and most successful attacks are simple.?
• AI has a Bigger attack surface … This resonates well with the 2023 RSA Innovation Sandbox winner, HiddenLayer.
• Human intelligence has been underestimated in the era of GPT4.? Best practice in cybersecurity should be AI augmenting Human, well resonates with RSA Innovation Sandbox startup AnChain.AI latest innovation: Web3SOC.ai , the AI powered Security Operations Platform for Web3 digital assets.

Topic 2: Rise of Generative AI and its Impact on Cybersecurity.

• Job security.?
• In the SOC team, What jobs would be replaced first ? T1 analysts? or T3? And why??
• Argument: Creativity skills in T3 analysts has been probably overestimated, as proven recently by the Generative AI.? Stability.ai , Midjourney can generate more creative arts than human artists. T3 analysts’ creativity skill in threat hunting might not be as unique as you think.?
• Job creation. Although AI will reduce the number of entry-level jobs in cyber security (specifically SOC Type 1 analysts), at the same time it will create more jobs to fight the new attack surfaces and privacy issues.?
• Code generation. GPT is generating insecure code, and with half of the code on github coming from copilot, that's a challenge worth meeting. But instead of being a reason to not use it, it also shows something amazing: junior+senior analysts & coders are recognizing it as an amazing upskilling tool. GPT4 may be on the level of the internet and youtube wrt an educational advance for teams, which may be the bigger takeaway.

Topic 3: AI safety and responsibility.

David Cass published his post on this event:

• AI has a regulatory problem that agencies have started to act on, especially around explainability & bias.?
• Unlike traditional deep learning, generative AI may be in a surprisingly better position as it typically generates natural language traces and, done right, is tied to source data that the responsible decision maker can examine.?
• AI has ways to go to be widely adopted and become a standard across cyber space with regulators still figuring out best ways of handling it and creating security / compliance frameworks around it;
• The GPT / LLM models in 2023 are fundamentally still an AI / ML model trained from human curated data. In the financial industry, there are published guidelines such as : SR 11-7 SUPERVISORY GUIDANCE ON MODEL RISK MANAGEMENT, by the Board of Governors of the Federal Reserve System Office of the Comptroller of the Currency.?
• ISACA , the largest organization for IT auditors, recently published machine learning white papers covered both technology and compliance risks. Download the white papers at ISACA.

Topic 4: Investment areas.

• It is increasingly difficult to pick early stage startups in AI security since everyone now will start embedding OpenAI models in their products.?
• Vertical-specific applications where founders have certain sets of skills to fine-tune models or access to data to train models will be key. For example, for Web3 security startup founders that come from tier 1 cybersecurity companies such as Mandiant, etc that have gained real world battle testing skills will be more likely to succeed.?

Lastly, thanks for all the top minds that contributed to this wonderful night! We have pushed the boundaries of AI x Cybersecurity Innovations!?

Follow me and stay tuned for more upcoming cool events!???????

Attendants list (partial):?

112 Capital, 451 Ventures, A16z, AnChain.AI, AWS, AGI House, Blumberg Capital, Cisco Research, Conviction VC, Crcm,? Dragos, Eastlink Capital, Exabeam, FSA Government of Japan, Galaxy, Generative Alpha AI, Graphistry, Inc. GSR, HP inc, Intel, OpenAI, Menlo Ventures, Race Capital, Resilience, SafeBase, SentinelOne, Shield Capital, Stanford, UC Berkeley, Oasis Labs.?

Disclaimer: This article was consolidated from various contributors’ collective notes and there is no guarantee of accuracy, nor should it be considered as financial or legal advice.