Where Is AI Going?

At the last World Economic Forum meeting at Davos, Prof. Yuval Noah Harari delivered an electrifying speech on AI. “No matter from which country you come, your country will soon face a severe identity crisis and also an immigration crisis,” he said. “They are likely to be loyal not to your country but to some corporation or government across the ocean most probably in one of only two countries, China or the USA.” He was referring not to human immigrants but to “AI immigrants”, who will be AI doctors or AI teachers.

Harari also held a “fireside chat” on superintelligence together with Prof. Max Tegmark from MIT, Chairman of the Future of Life Institute and an AI researcher. Harari defined superintelligence using the example of a financial system: an autonomous agent that could be released into the financial system to make one million dollars on its own. Tegmark defines superintelligence as non-biological intelligence that is vastly superior to humans across all cognitive processes.

Harari emphasizes that AI is fundamentally different from previous human inventions because it is an agent rather than a tool. While tools like the printing press or the atomic bomb require humans to decide how to use them, AI can make decisions and invent ideas independently.

There is a great sense of urgency in Harari’s words and tone. Tegmark does not disagree that there is a significant potential danger arising from the development of superintelligence, but he also states: “This is not inevitable. It’s a huge mistake I think many people make. That’s what’s going to happen,  as if we humans now are just some sort of passive bystanders, you know, eating popcorn, waiting for AI to take over the world.” His conclusion is quite dramatic: superintelligence will not be achieved because nobody wants it, especially governments of the superpowers which do not want to be overthrown by AIs.

In the meantime, we can and should be very excited about the power that AI brings. However, the level of large-scale adoption is relatively slow. For example, the U.S. Census Bureau shows that only about 10% to 11% of the manufacturing and retail trade sectors use AI. These results are based on questionnaires, and the specific question regarding AI is: In the last two weeks, did this business use Artificial Intelligence (AI) in any of its business functions? (Examples of AI: machine learning, natural language processing, virtual agents, voice recognition, etc.).

In Israel, there are many systems which could be dramatically improved by adopting advanced AI tools and agents. For example, visiting an Israeli hospital emergency department (ER) is an experience that often triggers thoughts about the gap between the image of the “Start-Up Nation” and real life. Sheba Medical Center launched an initiative called Project K which is, according to their website, “the world’s first AI-powered emergency department.” It will take much ingenuity, along with development, organizational and regulatory efforts, to truly integrate AI into an ER and achieve real productivity improvements ­­­— for the benefit of both patients and the hard-working, often exhausted ER staff.

Harari warns of superintelligence AIs taking control over humans. It seems more plausible that before using AI massively for good, such as improving healthcare or reducing the effects of climate change, humans will use AI to fight each other. As Harari states in 21 Lessons of the 21st Century: “We should never underestimate human stupidity”. There is a significant risk that humans will now possess powerful AIs that, by design or by mistake, could be aligned with human stupidity.