Novel Turing test was devised to assess AI’s capacity to generate $1 million
AI expert innovated fresh assessment to gauge future possibilities of neural network advancement
Mustafa Suleiman, a co-founder of DeepMind and an AI researcher, has introduced a novel iteration of the “Turing test” tailored for AI, in light of the progress made by neural networks in successfully passing conventional evaluations.
Suleiman has expressed that in this assessment, the goal is not solely to determine the machine’s intelligence, but rather to gauge its potential to exert a significant impact on the world.
The recently proposed AI test presents a challenging task: to generate $1 million on a web-based retail platform within a few months using a mere $100,000 investment.
The complexity of this endeavour, Suleiman emphasises, surpasses that of tasks like text or image recognition. Neural networks will need to actively interact with contractors, analyse research findings, and negotiate agreements.
Should AI achieve success in this challenge, Suleiman contends that it would unlock possibilities not only in formulating business strategies, but also in managing infrastructure, conducting political campaigns, and various other domains. He has termed this system “artificial capable intelligence” (ACI).
Suleiman envisions that the Turing test he has proposed will function as an early indicator for humanity regarding the trajectories of ACI development. He concludes by stating that ACI is poised to revolutionise nearly every facet of human existence, underscoring the need for proactive preparation for this transformation.
