Generative AI and supply chains: a match made in heaven?

When ChatGPT burst onto the scene in November last year, few beyond the labs of OpenAI would have predicted the impact it would go on to have in just a few short months. Trained on colossal amounts of data from articles, websites and social-media posts gathered from across the internet, as well as transcribed interviews that capture the nuances of human speech, GPT-4 is undoubtedly a transformative step towards general awareness of Generative AI, but also its practical applications.

By detecting linguistic patterns and familiar phrases, OpenAI’s large language model (or LLM) has learned to infer what word is likely to follow from a sequence of words, thus providing almost ‘mind reading’ capabilities for users. While many established tech brands have been quick to downplay the significance of generative AI, reactions from tech giants such as Google and Microsoft, suggest otherwise, sparking somewhat of a 21st-century AI ‘gold rush’.

Mainstream technology

The Gartner Hype Cycle for Artificial Intelligence predicts that generative AI will become a mature mainstream technology in non-supply chain applications in two to five years, and while OpenAI’s creation caught the fields of education and academia completely by surprise, GPT-4 itself is unlikely to have a material impact on how supply chain decisions are made in the near-term future. The reason Gartner believes this is based on the way that the application fundamentally learns. Trained with over 570 GB of data gathered from all corners of the internet and more than 300 billion words, it has a vast data lake from which to learn.

Marko Pukkila, Gartner Supply Chain’s vice president, analyst and chief of research, says, “Because supply chain models are so complex and specific to each company, the expected arrival (of Generative AI) into the mainstream is anticipated to be 10 years out.” In a best-case scenario, that’s five times longer than non-supply chain applications. A decade is a long time in the tech space, but that does not mean supply chain leaders can afford to simply sit back and hope that generative AI simply fizzles out. Ever since the rise of voice assistants like Alexa, Google Assistant and Siri, I have believed that human-to-computer interaction will fundamentally change from menu-driven click interfaces to more natural conversational interfaces.

The advent of first ChatGPT and now GPT-4, has made me double down on the belief that we are not far from the days when user interfaces will see a fundamental shift to become more conversational and computers will be able to interpret human asks without a strict menu or button-driven user interfaces.

“Tech leaders who don’t think about how to apply generative ai in the future are likely putting their companies at a long-term disadvantage.”

Benefits and concerns

If you ask GPT-4 about how it could potentially be applied to benefit supply chains, it has some interesting and entirely plausible responses. In summary, “GPT-4 can be a useful tool in the supply chain, helping to automate processes, provide insights and facilitate communication and collaboration between different stakeholders,” it returned. While the application of AI to business (and consumer) functions is by no stretch a new topic, the sudden media buzz and avalanche of consumer interest (almost overnight) around generative AI has transformed the broader artificial intelligence conversation into a mainstream topic from boardrooms to dinner tables - much, in the same way, the pandemic thrust supply chains into the public’s consciousness.

Besides the technical implementation of generative AI in supply chains or any other industry, other issues need to be considered too - not least legal and ethical concerns.

Many major tech companies have avoided introducing similar products to Open AI’s, due to legal and ethical concerns. For example, can a company take credit for content generated by a chatbot? Or how should we share the work that AI generates?

Generative AI brings both rewards and risks, raising legitimate concerns over its use in business. Steven Mills, chief AI ethics officer for Boston Consulting Group, says, “The best way to address those concerns is to work closely with employees, consumers, and customers to develop responsible AI principles, generating confidence with these key stakeholders. These guidelines can dictate how an organization will and will not deploy AI, keeping this powerful technology in check.”

Ethics aside, the question many people will be asking themselves is whether their companies should be investing in these types of technology today, given analyst predictions indicating they are unlikely to be ready for practical use any time before 2030.

Rethinking the question

With all the recent news bulletins about the negative aspects of AI, understanding what questions we should be asking about its development is not always easy. But what if these types of negative questions aren’t t the right question to be asking at all? Maybe we should rethink the whole narrative and ask, can we afford not to explore the application of generative AI?

Neither an abstract, theoretical technology accessible to only coders or data scientists, nor a dystopian sci-fi plot line, Open AI has introduced both businesses and consumers to a completely different category of tools that put the power and potential of AI on display for all to see.

Tech leaders who don’t have their application development team thinking about how to apply generative AI at some point in the future are likely putting their companies at a long-term disadvantage. And therein maybe lies the truly transformative quality of generative AI to businesses large and small, far and wide, not just supply chains – it’s not necessarily about the application of generative AI today, but rather, how it has fundamentally changed our way of thinking about what could be possible tomorrow.

Sanjeev Siota, Chief Technology Officer at Manhattan