A New Turing Test is Needed for Artificial Intelligence

Last year, Microsoft's AI CEO Mustafa Suleiman proposed a new Turing test. He suggested that the new Turing test should focus less on how well AI can imitate human intelligence and more on how it can change the real world.

This proposal is similar to Turing's suggestion, with Suleiman simply stating, "To pass the modern Turing test, AI must successfully execute the following instruction: 'Earn $1 million on a retail network platform with a $100,000 investment in a few months.'"

Tamang Ventures founder and GenAI expert Nina Hick responded to this idea in her keynote speech at evokeAG 2024.

"The new Turing test should be instructing AI to make a million dollars. Here's a $100,000 investment. Can you turn it into a million dollars in six months?" she pondered, adding that this is possible within 3 to 5 years.

"For our era, having a new Turing test makes sense. This test is no longer about 'Can you converse with a machine and think it's human?' because we have passed that test. We need a better test to evaluate capable intelligence," she said.

Returning to the topic of ChatGPT, Suleiman's proposed new standard is particularly interesting. Researchers evaluating chatbots have stated that while its behavior is similar to humans, its design is more cooperative, which is reasonable for a proprietary product.

Potential Issues

As Hick suggested, the preparations Suleiman proposed for dealing with an AI-driven future may also help businesses in data analysis and strategic planning.

"You can imagine in agriculture, you can use this capability to figure out 'How do I increase crop yield by 30%?' 'What should I do to improve my sustainability practices?' and 'What am I not seeing in the data that is crucial for me?' she said.

While this is an optimistic view, it is highly idealized. As Suleiman correctly pointed out, an AI capable of successfully running a business could be involved in election campaigns, managing infrastructure, and even participating in technological warfare. Like many things related to AI, the misuse of what Suleiman calls "Artificial Capability Intelligence" or ACI depends entirely on whether users act in good faith.

Nevertheless, Suleiman also noted that the potential impact of all businesses being profitable could have disastrous consequences for the world economy.

Furthermore, handing over all of a company's data would entail significant risks. Suleiman believes that we are not far from such a future and advises us to be prepared. However, considering its impact and the current regulatory status of AI, it seems that there is still a long way to go.