With the advent of ChatGPT everyone is talking about the future role of Artificial Intelligence (AI) - and a fast growing number of businesses have prioritized utilizing AI in their companies. If Hollywood movies have over dramatized the evil role AI could play in the world, then business leaders could be seen as overestimating the beneficial role it can play in their organizations. While the widespread use of generative AI (apps like ChatGPT) is new, AI itself was developed - and has been evolving in theory and in application - since the 1950s. <h3><b>Where AI Could Make the Biggest Impact</b></h3> Like early forms of automation, AI is on a path to take over repetitive tasks that humans are doing - and do them faster and more accurately. However, whereas automation was about programming a computer to do a predefined task and operate within predefined rules, AI is about creating computers or apps that can make their own decisions - albeit still based on human input. The long term vision is the development of software that can supersede human thinking capabilities. “Companies are investing a lot in visibility. One of the big lessons from the pandemic was that to reduce supply chain disruption, you need a very sharp picture of what’s going on throughout the entire chain. The faster you can access this information the faster and more accurately you can react. This could be finding other suppliers or something as simple as notifying your customers that a shipment is not coming,” said Yossi Sheffi, Director of the Department of Logistics and Transportation at MIT university (USA). “There’s a lot of big investment in warehouse robotics, which is becoming one of the largest robotic applications in the world. Many are going beyond automating and are looking to AI infused robotics,” he explained. When you take the old conveyor belt and add robotics you can make it faster and more reliable. However when you add AI, you can achieve greater flexibility. “When you use robots with AI that can move around the warehouse, they can respond to changes in demand and be much more flexible. So this is a huge area of investment,” said Dr. Sheffi. Another area of investment is in risk analysis. For supply chains this could let brands know if any of their global suppliers is at risk. “Financial reports or credit reports only show you the past. Using Large Language Models, we can now sift through social media to pick up signals that there could be problems. Things like failed mergers, late payments or late deliveries are alerts that there's some problems, or that the executives are not paying attention to the business,” he said. <h3><b>New Systems to Meet New Demands</b></h3> Success with AI comes down to looking at systems and models in a new way, not merely automating the old way. One core area where AI is playing an important, but less glamorous role, is in what’s called ‘hyperautomation’. “Hyperautomation is just automating fast. Every company has this ratio of human tasks to automated tasks. We all think that that ratio's gonna be static and that the more we automate the fewer human tasks will be required in the company. “But history's showing us that the more we automate, the more human tasks we invent and that ratio doesn't seem to be closing the gap. It seems like we are creating new jobs for humans to do as we automate the old ones. For example, the cappuccino machine in your house was the beginning of baristas and Starbucks. Not the end of it,” said Robb Wilson, founder, lead designer, and chief technologist behind OneReach.ai. OneReach.ai was the highest-scoring company in Gartner’s first Critical Capabilities for Enterprise Conversational AI Platforms report. AI is also facilitating digital twin technology by helping it to analyze huge amounts of data faster. “People are looking at reconfiguring supply networks almost on the fly. But to do this, you need to know all the suppliers around the world who can make this product, not only who made this specific product. Even those who have the equipment to make this product - and in a pinch - can make it. Who are their suppliers? You look at the whole supply chain using data that we couldn’t access before. And then it comes down to analyzing it quickly and quickly making sense of it,” said Dr. Sheffi. <h3><b>Where AI Has Weaknesses</b></h3> At present AI is dependent upon massive amounts of data collated by human beings, and thus has a natural bias. AI is being ‘trained’ based on the past, making it less equipped to deal with the future. “It doesn't matter if it's based on linear regression, or machine learning or whatever. All forecasting is based on past patterns. When the pandemic hit, past patterns had nothing to do with new patterns,” said Dr. Sheffi. “So people immediately jump in and start doing things by hand and start calling other people and talking to customers and suppliers just to get an idea of what was a reasonable forecast,” he said. All software and hardware can be victims of malware, cyberattacks or even physical attacks or other disasters (fires, power outages, weather). This is where its people who end up saving the day. Another challenge for companies is to distinguish between what AI can do and what is feasible within their organizations. “AI can do a lot more than most people think it can. However the focus should be on feasibility. Just because AI can do something, doesn't mean it will work for your business, and companies. It's so complicated and hard to implement that right. Most companies can't do this. So they need to consider what is feasible for them to implement,” said Mr. Wilson. One way to test the viability of certain apps or software is to ask the developer or seller of the software whether they use that app in their own company. <h3><b>Why AI Needs Humans </b></h3> Yet, in the post Covid era, companies are racing to digitize and automate their operations “COVID gave us one picture of when automated systems don't work. But automated systems are also subjected to cyber attacks, they are subjected to all kinds of malicious things. “Now we have the whole generative AI that is really not replacing people, but actually supporting them,” said Dr. Sheffi. “The main thing that people have is an understanding of the context system. it's not clear that even GPT 456 will have this ability. Context could come in with regard to moral judgment. It can come in understanding risk management. For example, we may have machines that can suggest several several courses of action. Now someone needs to make a decision on which course of action to take. The question is, ‘now I have to make the decision, ‘are we going into recession or not? What's the mood in Washington?, Is China gonna attack Taiwan? and so forth.” A lot of these are human judgments. Empathy and moral code come under the title of context. So the best combination is people working with machines. “The question is we have to know when to use it and how well to use it to judge the results. It can generate very accurate results and it can also generate nonsense,” he added. Thus, it comes down to people making the final judgment call on what AI has generated. <h3><b>Good But Not Perfect</b></h3> However, “this automation remains far from human intelligence in the strict sense, which makes the name open to criticism by some experts. The ultimate stage of their research (a "strong" AI, i.e. the ability to contextualize very different specialized problems in a totally autonomous way) is absolutely not comparable to current achievements ("weak" or "moderate" AIs, extremely efficient in their training field). The "strong" AI, which has only yet materialized in science fiction, would require advances in basic research (not just performance improvements) to be able to model the world as a whole,” per the Council of Europe. “I am of the opinion that the new AI has a chance of bringing an unprecedented era of prosperity, of incredible efficiency, and incredible productivity. We just have to make sure that it doesn't fall into the bad side of what it could do,” said Dr. Sheffi. He explains that in the early days of the internet when everyone saw it as magically connecting people worldwide and bringing forth peace and harmony, it ended up being used for both good and bad. However, he points out that in these early days of advanced AI there is a lot of discussion at both the developer level as well as within governments and other organizations about creating guard rails to prevent the malicious use of this technology.