Carlos Sentís delivers two sessions on AI for National Security at the Naif Arab University of Security Sciences, Saudi Arabia

🔸 AI for National Security: Naif Arab University of Security Sciences, Saudi Arabia.
🔻 AI Limitations and Risks

🔹This year I had the distinct honor to address National Security leaders in Riyadh to discuss how AI can be used by bad actors to harm Governments, Society and Critical Infrastructure, as well as how it can be used by the Police, Military, Foreign and Interior Services to prevent attacks and mitigate risks.

🔻 AI Limitations and Risks:

🔹AI and any other technology can and will often fail, it still has and will continue to have limitations, yet its power and potential could be the greatest any invention has ever brought
🔹Despite efforts and investments AI Alignment is not possible. We must assume the following AI Risks will remain constant and ever more critical:

🔹AI limitations in its outcomes, errors Algorithmic biases and hallucinations
🔹AI technical system errors or control failures
🔹Intentional Bad actors using AI or AGI against others
🔹Hijacking or hacking AI systems, including model poisoning, model injection, jailbreaking and other
🔹Billions of AI Agents or AGI Agents operating autonomously to attack systems, creating new attacks or attacking with multiple different approaches
🔹Leaks or re-training on confidential information submission to AI tools or leaks
🔹AI over dependence
🔹Potential escalation of Autonomous Weapon Systems or AI - Automation driven weapons in seconds akin to trading bot market collapses as algorithmic responses to false triggers
🔹Super AGIs overpowering humans in areas of the digital and physical world
🔹AGI acting against society or specific groups independently being able to leverage millions of times more Information and intelligence processing needs

Will continue to share some of my research findings around the topic, and how it specifically affects large international organizations, government institutions, enterprises and professionals.

My deepest appreciation to Dr Adel El-Adawy, high level advisor and researcher on the Intersection of Emerging Technologies, National Security and International Relations, for his generous invitation. Dr Adel wrote a fantastic paper about Artificial Super Intelligence that I recommend everyone reading.