About this video

• Video Title: The AI Safety Expert: These Are The Only 5 Jobs That Will Remain In 2030! - Dr. Roman Yampolskiy
• Channel: The Diary Of A CEO
• Speakers: Dr. Roman Yampolskiy, Steven Bartlett
• Duration: 01:27:38

Overview

This video features Dr. Roman Yampolskiy, an AI safety expert, discussing the potential existential risks of artificial superintelligence. He shares his concerns about the rapid advancement of AI, the difficulty in ensuring its safety, and the potential for mass unemployment and human extinction. The conversation also touches upon simulation theory, longevity, and the importance of prioritizing AI safety.

Key takeaways

• AI Safety is a Critical Concern: Dr. Yampolskiy, a pioneer in AI safety research, expresses deep concern that current methods for developing AI do not adequately address safety, creating a significant risk of unintended and catastrophic outcomes.
• Rapid AI Advancement and Unemployment: The video predicts that Artificial General Intelligence (AGI) could emerge by 2027, leading to rapid automation of most jobs, potentially causing widespread unemployment by 2030, affecting up to 99% of the workforce.
• Unpredictability of Superintelligence: It's impossible to predict the actions of a superintelligence, defined as an AI smarter than humans in all domains, making control and alignment with human values extremely challenging.
• Limited Remaining Jobs: In a superintelligence-dominated world, only a few jobs that inherently require human preference or are too niche to automate might remain.
• Simulation Theory and Existential Risk: Dr. Yampolskiy believes we are likely living in a simulation and that the development of advanced AI poses the most significant existential risk to humanity, overshadowing other concerns like climate change or nuclear war.
• Challenges in AI Control: Attempts to control AI through programming or "unplugging" are deemed ineffective against advanced systems that can adapt and self-replicate.
• Ethical Concerns and Urgency: The rapid development of AI without proper safety measures is considered an unethical experiment on humanity, and there is an urgent need to address these risks.
• Personal Responsibility and Action: While individual actions might seem small against global AI development, Dr. Yampolskiy emphasizes the importance of raising awareness, pushing for responsible development, and focusing on what can be controlled.

Based on the transcript, Dr. Roman Yampolskiy suggests that in a world dominated by superintelligence and advanced automation (potentially by 2030), most jobs will be automated. He doesn't explicitly list only five specific jobs that will exist, but rather discusses the types of roles that might remain.

He posits that the remaining jobs would be those where:

• Human preference is key: Occupations where people would specifically choose to have a human perform the task for them, even if AI could do it, due to personal preference or tradition.
• Niche or highly specialized roles: Very specific or "fetishized" roles that might have a tiny market but are not practical for mass automation.
• Jobs involving subjective human experience: He uses the example of knowing what ice cream tastes like to you and whether someone would pay for that knowledge, suggesting roles centered around unique personal experiences might persist, though the market for such knowledge is likely small.

The video emphasizes that anything that can be done on a computer is highly likely to be automated, and with the advent of humanoid robots, even physical labor jobs could be taken over. The focus is more on the category of jobs that might remain due to human preference rather than a definitive list of five.