Unlocking the Complexities of AI Security and Standards - Insights from David Snyder for the World Technology Summit

In his recent Manifesto, “The Intelligence Age”, OpenAI’s Sam Altman paints a bright future. I am a technology enthusiast myself, and agree, mostly. Sam states:

“If we want to put AI into the hands of as many people as possible, we need to drive down the cost of compute and make it abundant (which requires lots of energy and chips). If we don’t build enough infrastructure, AI will be a very limited resource that wars get fought over and that becomes mostly a tool for rich people.”

That criticality of AI Infrastructure is exactly why I agreed to get involved in the organization of the IEEE World Technology Summit focused AI Infrastructure (WTS) in November. Security is a key concern.

Recently, Adam Drobot, Chairman of the Board, OpenTechWorks, Inc., and one of the co-organizers of the WTS, discussed the critical intersection of cybersecurity and AI with David Snyder, a cybersecurity advisor, who is chairing the security section at WTS.

You can still reserve your spot at WTS here, and you can find the full video of Adam's and David's discussion here:

David’s experience gives him unique insights into the security challenges posed by AI. I especially liked his Three Dimensions of AI Security - Security of AI, Security from AI, and Security by AI.

In David’s mind, "Security of AI" refers to protecting AI models and their underlying infrastructure from malicious attacks. "Security from AI" involves preventing AI from being used to develop and deliver malicious content, such as disinformation. Finally, "Security by AI" pertains to using AI to defend against cyberattacks. These three dimensions underscore the multifaceted nature of AI security and the need for comprehensive strategies to address each aspect.

Another topic that David and Adam touch on in their discussion is The Role of Data in AI Security, as the vast volumes of data required for AI add a new layer of complexity to security concerns. David channels his inner Spiderman by emphasizing that “with great complexity comes great responsibility” as large datasets and complex models can introduce concerns around bias, equity, transparency, trustworthiness, and privacy.

In addition, David stresses the Environmental Impacts of AI. The substantial power consumption of AI systems, coupled with the need for cooling solutions, has significant environmental impacts.

With hundreds of proposed laws and regulations worldwide, compliance becomes a significant burden for organizations. Addressing the Regulatory Landscape, David advocates for performance-based requirements, which focus on desired outcomes rather than prescriptive rules, to strike a balance between regulation and innovation.

Standards and Interoperability play a crucial role in ensuring interoperability and security in the AI and ML space. The work of the IEEE and the International Standards Organization (ISO) in developing standards for AI is critical, and there also guidance coming from the executive branch in the United States, as national legislation faces challenges. Standards help create a common framework for AI development and deployment, fostering trust and reliability in AI systems.

The World Technology Summit is coming up fast. You can still reserve your here, and I am looking forward to seeing you in November.