Artificial Intelligence: 

Time Magazine recently wrote, “In many ways, 2023 was the year that people began to understand what Artificial Intelligence (AI) really is – and what it can do.”  What is Artificial Intelligence? How is it changing our world? Where does the U.S. stand in the global AI arena? And how does Israel fit in?

​

AI, in a nutshell, means giving computers the capability to learn and think like humans. Until AI, computers could perform well-defined tasks, such as “Calculate this spreadsheet” or “Tell me what the weather will be like this weekend.” The computer would perform a large number of calculations in a short period of time and then regurgitate the answer. AI is a qualitatively different way of doing business. AI can broadly be divided into categories: 

• Weak AI, which performs specific tasks,

• Strong (or generative) AI, which has the ability to learn, think and perform new activities like humans. 

​

Weak AI is very good at “classification.” It can analyze a picture and find all the dogs. It can look at a CT scan and find potential pre-carcinogenic symptoms. Using cameras, laser and radar data, Weak AI can determine when it is safe for a car to enter a roundabout. According to many experts, self-driving (autonomous) cars should be ubiquitous within a decade. Weak AI is powered by a machine learning algorithm called “deep learning,” which uses millions of examples to train a “neural network,” a machine-learning algorithm. Neural networks have been around for more than 65 years, but only recently has there been sufficient data and computing power to train a neural network so that it can be useful. Neural networks learn by themselves but they need to be “tweaked” by humans. For example, we are all familiar with the “Captcha” in which we must prove to the computer that we are really humans. We are given a picture divided into 16 squares and we must click on the squares that have traffic lights. What we are doing is helping Google Maps train its neural network. Three of those squares contain objects that are obviously traffic lights, and one square contains an object that might be a traffic light. We prove that we are human by identifying the three obvious traffic lights and then we help out Google with the fourth object. When Google finally releases a fleet of autonomous cars, we can say that, in some small way, we helped them out. 

 

Strong – generative – AI is a giant leap from Weak AI. While generative AI, also known as “GenAI,” also uses deep learning, the product of GenAI is seminally different than the product of Weak AI. Weak AI will tell you if that thing in the corner of the picture is a traffic light. GenAI will draw you an oil painting of a traffic light with a Salvadore Dali motif. The best example of this kind of generative AI – or “text-to-image system” – is called “DALL-E 3.” According to its inventors, the OpenAI Group, “DALL·E 3 understands significantly more nuance and detail than our previous systems, allowing a person to easily translate his ideas into exceptionally accurate images.” Its ability to create images is uncanny and sometimes downright creepy, such as “Darth Vader ice fishing” or “Ronald MacDonald performing open-heart surgery.” One highly problematic byproduct of GenAI image creation is called “deep fakes” in which the faces and voices of one person are swapped for another. Deep fakes can create entirely original content where a person is clearly shown doing or saying something they did not do or say. If this sounds reminiscent of Forrest Gump, it should be. In Russian President Valdimir Putin’s 2024 New Year’s address, his neck seems, well, wrong (see for yourself), leading people to speculate that the entire address was a deep fake and that Putin was incapacitated. 

​

By far, the most well-known example of GenAI is ChatGPT (also integrated into and/or copied by Google Bard and Microsoft Copilot). ChatGPT, a product of OpenAI, is a Large Language Model (LLM), a type of GenAI algorithm that uses deep learning techniques and massively large data sets to understand, summarize, generate, and predict new content. ChatGPT has been used to write doctoral dissertations, prepare legal briefs, and to determine what is the very best birthday present to buy my wife this year. More importantly, ChatGPT can carry on conversations in a way that replicates human conversation, and by doing so, has broken the Turing Test. The Turing Test, proposed by Alan Turing in 1950, is a straightforward method of determining whether a machine can demonstrate human intelligence: If a machine can engage in a conversation with a human without being detected as a machine, it has demonstrated “human intelligence.” ChatGPT passes the Turing Test with flying colors, requiring a new method for assessing AI. The newest version of ChatGPT, called GPT-4o, can carry on voice conversations with users in near real-time, exhibiting human-like personality and behavior.

​

Now that we know what AI is, we can begin to understand why it is so critical for the U.S. to excel at AI. Unsurprisingly, the U.S. and China are engaged in a war for global AI dominance. Prima facie, AI does not seem to be a commodity with limited availability. Chinese AI prowess should not preclude American prowess – there is more than enough AI to go around. What, then, does “winning the AI battle” entail? To start, the premise of there being sufficient AI to go around is only partially true. AI will have a profound impact on state power, mainly through increasing economic growth and enhanced military capability. Additionally, global leaders in AI will set norms around its use and standards that will be used in its implementation (particularly data collection that could infringe on personal privacy). China has already made great strides, using the data of more than one billion citizens – with negligible rights to privacy – to train its AI applications. If China “wins this battle,” the entire world could end up “doing AI” the Chinese way, which may not be amenable to American standards that are already proliferating in American companies. To this end, on February 11, 2019, President Donald J. Trump launched the American Artificial Intelligence Initiative, the U.S. national strategy for maintaining American leadership in AI, by signing Executive Order 13859. The initiative contains 6 components:

1. Invest in AI research and development,

2. Unleash AI resources including data and models,

3. Remove barriers to AI innovation,

4. Promote an international environment supportive of American AI innovation,

5. Embrace trustworthy AI for government services and missions, and

6. Train an AI-ready workforce.

 

Item (4) makes a critical statement: The U.S. cannot go at it alone. U.S. AI leadership requires collaboration with its allies. This is where Israel comes in. Israel’s AI ecosystem is undergoing explosive growth, characterized by a major influx of research centers, establishments, and talent, forming what could be considered one of the global leaders of AI over the next decade. Israel is a hotbed for AI talent, tracking at more than 4000 developers, engineers, and data scientists working on AI-related research, development, and integration. All of this in a country of about 9 million. The Israeli Ministry of Science has designated AI as a “vital area of research” and in 2019 began offering financial incentives to encourage academic AI research.  Between 2013-2022, Israel has the fourth highest number of AI startups (402) and amount of private investment ($11B) in the world, trailing only the U.S., China, and the U.K. 

​

The U.S. already recognizes Israel’s AI prowess and has moved to benefit from it. The Israeli Mobileye company produces a system that teaches cars to visually recognize the environment in which it drives. It ensures safe driving even at high speeds, avoiding both vehicles and pedestrians, and the system sounds an alarm if the car veers out of its lane. Mobileye technology is found on millions of cars around the world. In 2017, Mobileye was purchased by the U.S. behemoth, Intel, for more than 15 billion dollars and Intel has relocated its autonomous car division to Israel. To feed off the Israeli AI tidal wave, many more leading U.S. companies have either moved their AI Research Centers to Israel or have opened up new ones in Israel. These companies include General Motors, Nvidia (the world leader in manufacturing AI computer chips), Microsoft, and Google, who has opened two Israeli research centers working on machine learning, natural language comprehension, and machine perception. Collaborative academic research between the two countries is also booming. In the past 5 years, the U.S. – Israel Binational Science Foundation has funded no less than 28 collaborative research programs in the field of AI and another 45 in the field of Machine Learning.  

​

Under the auspices of the Israel Innovation Authority, Israel is using  AI to forward its leadership in “bioconvergence,” which fuses life sciences with an array of technologies from fields such as mathematics, engineering, and physical and computational sciences, creating a total far greater than the sum of its parts. AI is a key enabler in bioconvergence. It is unsurprising that many Israeli startups specialize in the field of AI and life sciences. Here are a few notable examples:​​​​

​

• BrainQ uses non-invasive, frequency-tuned extremely low frequency and low intensity electromagnetic fields (ELF-EMF) to promote neurological recovery in the central nervous system. The AI-powered device tailors the electromagnetic field characteristics to suit each patient.​​