Ads Dawson, a staff AI security researcher, explores the evolving landscape of artificial intelligence (AI), from its foundational role in society to its potential as a security threat. He reflects on AI milestones like IBM Watson's Jeopardy victory and AlphaGo's defeat of Go masters, illustrating AI's capacity to handle complex tasks traditionally seen as uniquely human. Dawson highlights the vulnerabilities of narrow AI systems to hacking, such as adversarial attacks, and discusses the emerging risks associated with artificial general intelligence (AGI), likening them to over-eager interns capable of introducing security flaws through misuse. As companies increasingly integrate large language models (LLMs) into various processes, Dawson warns of security risks when these systems are not adapted to AI's limitations. He speculates on the distant yet significant implications of artificial super intelligence (ASI), a concept currently relegated to science fiction but crucial for shaping risk awareness. Dawson encourages a proactive approach to AI security, advocating for continuous testing and exploration of AI vulnerabilities to prevent future threats. Through his engaging narrative, he underscores the importance of hackers in identifying and mitigating AI risks, ultimately aiming to keep AI development secure and beneficial.