The researchers found that although the results achieved by AI models rose as they gained more computing power, AIs still performed poorly on most tasks compared with humans. Artificial intelligence (AI) researchers at 132 institutions, including Google, have proposed an update to the Turing test for rating the capabilities of AI technology. The researchers suggested using the Beyond the Imitation Game benchmark (BIG-bench), comprised of 204 diverse tasks designed not to be fully solved by state-of-the-art AI models. The researchers observed that AI models performed poorly on most of the tasks compared with humans, although they showed improvement as computing power increased. Adrian Hilton of the U.K.'s University of Surrey said, "I can see that setting up a set of benchmarks is one way of comparing one machine-learning algorithm with another, one AI with another. But I don't think that necessarily answers the question of intelligence though."

More than 400 AI researchers, including several from Google, have proposed an update to the famous Turing Test to allow us to rate the capabilities of AI technology. AI language models have become surprisingly, and often shockingly good at conversing with humans in recent years due to ever larger scale: more computing power and vast sets of training data.

Additional operational enablement can be achieved by using technologies like AI to improve how operational and technology teams engage with security. For example, using technology available today, the time required to complete routine security processes can be reduced significantly by using AI to automate resource- or time-intensive aspects of these processes. For operational teams, improvements in security process efficiency reduce the friction associated with following security requirements. Developments in AI technology are expected to unlock more opportunities to improve cybersecurity operations and support the balance of risk and return.