Since it was launched in April the new iteration of General Electric (that is, GE Aerospace) has done an impressive job of conveying to the world that it is a manufacturing business. It’s headquarters is a production center, not a downtown office tower. It’s staffing up with engineers and technicians, and investing to strengthen the network of suppliers supporting its production of propulsion and industrial power systems. Then, last month GE let slip some news that shows it is not immune to the trends that sweep through the work of every business and the life of every individual.
I’m referring to the world’s fascination with artificial intelligence. Working with Microsoft’s Azur AI, GE reports it created a private artificial intelligence network for secure use by its 52,000 employees. The manufacturer reports that since June the custom AI Wingmate tool has received approximately 500,000 queries, searches, and other employee interactions, and generated more than 200,000 pages of activity from different AI chats.
“Generative AI is changing the way we all work, enabling us to be more productive with our daily work tasks than ever before,” GE’s chief information officer David Burns said. “Ultimately, this will strengthen our ability to help the aerospace industry meet our highest objectives of enhanced safety, sustainability, and reducing supply chain constraints.”
That’s certainly one explanation for the company’s embrace of AI, the digital phenomenon that during the past two years has floated like a fog into the ether of public and private activity, changing how we see things, and the atmosphere from which we draw ideas.
For centuries, ‘originality’ has been the highest objective for intellectuals, scientists, and artists seeking to advance understanding. Originality also became a highly marketable quality in commercial activity. The prerogative of originality is not so much to break rules or standards (that’s called iconoclasm) but rather to draw out of established bases of knowledge some new insight or interpretation that would make existing information seem novel or more accessible.
The premise of AI inverts this in some way. It’s not testing out new ideas but sifting through existing information on the premise that carefully evaluating all of it will discover patterns – and that those patterns are proofs that we can trust in place of our own discovery, or consideration, or judgement.
The promise of AI has been much more – it’s a drafting tool, a calculator, a simulator. Somehow or other the answer to every business dilemma. It’s a shortcut. And AI’s parallel function as a creative platform for alternate realities makes it a time-suck – not only an engaging and responsive interlocutor, but one that gratifies users who are encouraged to feel that their random impulses are validated as completed work.
Now, GE’s explanation that its workers are using AI to conduct advanced searches that support project work is quite different than someone wasting hours generating graphics of superhero cats – but it’s also not true that AI will be redesigning better engine blisks or fuel nozzles. That will be the work of engineers, and producing components like that is the work of manufacturers.
Manufacturers and suppliers across several industries that have discussed AI with me eventually but invariably come to a similar point: Why all the enthusiasm? One engineer told me it’s a "buzz word."
They know that this novel new development rests on the availability and reliability of huge volumes of data – data such as manufacturers have been collecting and collating for decades. The task of reading that data is daunting, and new tools that speed and simplify the reading are welcome.
But they also know that the quality and propriety of the data is critical to the reading, which makes them cautious about what they share and the origins of the data they consider. They want to make certain that they do not become the product of someone else’s AI project.
And, ultimately, they know that the work of manufacturing is not really about answering questions; it’s about setting targets for results and solving problems based on conditions and variables.
Manufacturers can recognize how time can be wasted on fruitless tasks, just as they know that collected data is a starting point to solving current or anticipated problems. And regarding artificial intelligence, they recognize that we need to think this through.