AI’s Role in Escaping the Simulation Hypothesis

AI security expert and author of the book “AI: Unexplainable, Unpredictable, Uncontrollable,” Roman Yampolskiy, has assessed the probability that humanity lives in a simulation as one hundred percent. In his view, AI could help confirm this and find a way out.

In his article “How to Escape the Simulation”, the expert noted that superintelligent artificial intelligence could help overcome the limitations of our world.

“We used ‘boxing’ for AI as a potential tool to control it and realized that the neural network will always escape. This skill can be used to get out of the virtual box if we are in one,” said Yampolskiy.

Yampolskiy pointed out that the intelligence of our simulators will play a key role in this process.

“If the simulators are significantly smarter than us and the superintelligence we created, they can probably control us, as higher intelligence can manage lower intelligence, at least for a while. However, if our superintelligence, despite limited local resources, somehow manages to rise two levels higher, it might succeed,” he noted.

Yampolskiy suggested one way to escape the simulation. According to him, we need to create a simulated copy of our universe, place AI in it, and observe how it escapes. We could then replicate the neural network’s method or join it.

Besides his thoughts on simulations, the expert also noted that AI will lead to the extinction of humanity within the next 100 years with a probability of 99.9%. He did not specify reasons for this scenario, only stating that it is inevitable because superintelligence will begin to possess self-awareness.

Earlier, Near Protocol co-founder Ilya Polosukhin stated that AI will not destroy humanity unless programmed to do so.

Back in May, Elon Musk identified artificial intelligence as his primary concern.