![](https://ainews.one/wp-content/uploads/2024/06/0x0-802x900.jpg)
The evolution of AI will continue to bring promise and challenges for manufacturing companies
These days, we’re never far away from a story about artificial intelligence (AI). Yet, in truth, it isn’t new. Even as far back as the 1950s, AI was helping create rudimentary problem-solving models and, throughout every decade since, the length and breadth of its use cases have steadily increased.
What is new, of course, is the speed with which AI’s capabilities have suddenly exploded in the last couple of years. With generative AI, tasks that once took humans hours or even days to complete are being executed within minutes by ever more sophisticated applications.
Eye-catching as this rapid progress has been, however, the real driving force behind AI isn’t AI itself; it’s the technology that supports and enables it. Just as the problem-solving models in the 1950s couldn’t have functioned without large computer mainframes, today’s AI wouldn’t exist without recent advancements in cellular technology, the cloud and high-density data storage.
As Sook Chua, an EY technology strategist, recently put it: “AI stands on the shoulders of emerging technologies and infrastructure, not the other way around.”
Inspiring but still limited
Manufacturers must therefore resist the temptation to get carried away by the hype surrounding AI, focusing instead on how it can be used to solve their real business challenges — both today and tomorrow.
Already, there are plenty of use cases to choose from, with most of them focused on boosting the productivity of human workers. It includes everything from automating inventory management and smartening logistics planning to fostering better communications and data sharing between companies and their suppliers.
Many firms are deploying AI in content creation, too, using it to develop text, videos and graphics more quickly and cost-effectively than ever. Consider, for example, how much more efficient it is to have AI develop a set of multilingual standard operating procedures than it is to ask a team of employees to work on writing and translating them for several days.
At the same time as being inspired by AI’s potential, the industry should also have its eyes open to the things it can’t do — namely, making recommendations and decisions. On the shop floor, for example, AI may accurately highlight hazardous working conditions or suboptimal processes, but it can’t — yet — think critically about how to resolve them.
Committee decisions
This leaves manufacturers with two jobs to do. First, they must understand the possibilities and limitations of AI, setting themselves up to harness the latest applications right here, right now. But at the same time, they need to be planning and preparing for what comes next.
It’s a considerable challenge. So, rather than lean on a single chief information officer or chief technology officer to tackle it, firms should establish a governance committee that acts as the custodian of the organization’s entire technological ecosystem, including the responsible development and deployment of AI.
Reporting to the CEO and interacting regularly with the board, this committee should be truly multidisciplinary, ideally comprising representatives from HR; legal; marketing; finance; and technology — data, cyber, infrastructure and IT risks.
Crucially, it must view its responsibilities as stretching well beyond just the application of AI, taking the lead in ensuring all technological investments are aligned with the organization’s overall business priorities and goals.
This includes continually monitoring and evolving the strategy in line with advancements in the technology infrastructure itself. After all, as Sook Chua explains: “The more tools become available, the more quickly we’ll move into AI’s next stage.”
Now and next
Predicting exactly what this next stage looks like can be difficult, not least because recent progress has been so fast. Already, though, developments in areas like quantum computing and causal- or explanation-based AI are giving us glimpses of a world with greater speed and higher accuracy of training AI models, as well as AI reasoning and interacting in far more human ways, respectively.
Besides, even if uncertainty may still exist around the precise nature of AI’s tomorrow, one thing we can be sure of is that its capabilities will never be as limited or as inadequate as they are today. The AI era can offer manufacturers a future of near limitless possibilities — but only if they have a technology infrastructure that’s ready and able to unlock it.
The views reflected in this article are the views of the author and do not necessarily reflect the views of Ernst & Young LLP or other members of the global EY organization.
###