The auto company has trained more than 1,000 employees over the last year on a platform housing home-grown and purchased AI development tools, said Gil Gur Arie, Ford’s chief data and analytics officer.
Ford has long used AI throughout its organization, including in its plants, where AI analyzes data from equipment sensors for preventive maintenance; on its assembly lines, where smart robots build cars; and across the business to help manage parts inventory and other activities. In 2017 it acquired a stake in startup Argo AI as part of its efforts to develop an autonomous car.
With complex AI applications, such as autonomous driving, left to professional data scientists and other AI engineers, those trained on the new AI tools are looking to new areas to deploy the technology, said Mr. Gur Arie, a retired Israeli military intelligence corps colonel who joined Ford in May 2020.
“In other cases where it’s simpler, we can democratize that power,” he said.
Ford last week reported a steep drop in third-quarter profit as the computer-chip shortage dented factory output, although the company said supply-chain disruptions should slowly improve in the fourth quarter and throughout next year. Still, Ford said, dealership lots will be nearly bare well into 2022.
To help lessen the pain, Ford’s AI builders are working on an AI-optimization model that will help the company decide which vehicles should be shipped to which European countries so that car inventory is optimized to maximize sales, according to Ford. The model takes into account thousands of variables, including the carbon-dioxide emissions of each vehicle type, each countries’ emission standards, the amount of miles citizens in a particular country drive, as well as the adoption of electric vehicles and the size of vehicles preferred in each country. Ford said the number of variables being analyzed requires the use of AI, which is designed to handle large data sets.
Other applications developed through the program include an app to discover air duct forms that make car cabins quieter as well as a machine-learning model, built by a logistics team, designed to streamline the shipments of parts from warehouses to its more than 60 plants around the world.
“It’s a pretty complicated problem—even though you think it may be quite simple,” said Mr. Gur Arie, who said the logistics application is saving the company tens of millions of dollars in shipping costs.
Ford hopes that opening up AI development to a broader range of employees can significantly reduce the average time it takes to develop many applications, in some cases from months to weeks and even days.
Philipp Kampshoff, a senior McKinsey & Co. partner who leads the firm’s Center for Future Mobility, said car makers could unlock billions of dollars in savings with the automation and analytics provided by artificial intelligence if they can figure out a way to scale the technology.
One way to do that is by expanding access to AI development tools and training beyond data scientists and other AI experts, he said, a movement known broadly as AI democratization.
Some 15% to 20% of businesses that have embraced artificial intelligence have initiated AI democratization efforts, said Ritu Jyoti, group vice president of worldwide artificial intelligence and automation research at International Data Corp.
There are challenges with democratizing AI, Ms. Jyoti said. The models developed by these AI builders, for instance, need to be validated and tested thoroughly to avoid unintended results. Controls and governance also need to be put on the data resources and models these AI builders, or citizen data scientists as they’re often called, tap and use to ensure data integrity and appropriate validation metrics such as accuracy, fairness and explainability, she said.
Ford said it has measures in place. It puts guardrails around its data resources. It set up testing and validation procedures for all models developed by its AI builders. And it sometimes pairs its AI builders with its top data scientists on more complex applications to ensure quality.
Ford’s AI builders are using a platform that includes among other tools Domino Data Lab Inc.’s machine-learning operations platform, which provides a central hub for tapping compute resources as well as for cataloging and tracking models once they’re deployed, and an internally developed “feature store,” or repository, of reusable artificial-intelligence building blocks.
Subscribe to Mint Newsletters * Enter a valid email * Thank you for subscribing to our newsletter.
Never miss a story! Stay connected and informed with Mint.
our App Now!!