Ford uses non-IT professionals to broaden its AI expertise

Ford’s chief data and analytics officer Gil Gur Arie said the auto company trained more than 1,000 employees over the past year on a platform with AI development tools developed and purchased in-house.

Ford has long used AI throughout its organization, including at its plants, where AI analyzes data from equipment sensors for preventive maintenance; on its assembly lines, where smart robots build cars; And to help manage parts inventory and other activities throughout the business. In 2017 it acquired a stake in startup Argo AI as part of its efforts to develop autonomous cars.

With complex AI applications such as autonomous driving, left to professional data scientists and other AI engineers trained on new AI tools, they are looking for new areas to deploy the technology, Mr. Gur Ari, said a retired Israeli Military Intelligence Corps colonel. Who joined Ford in May 2020.

“In other cases where it’s easier, we can democratize that power,” he said.

Ford last week reported a sharp drop in third-quarter profits as computer-chip shortages hit factory production, though the company said supply-chain disruptions were gradually improving in the fourth quarter and through the next year. Should be. Still, Ford said, dealership lots will be nearly bare in 2022.

To help ease 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 European countries to maximize car inventory, according to Ford. to be adapted for sale. The model takes into account thousands of variables, including each vehicle type’s carbon-dioxide emissions, each country’s emissions standards, the amount of miles citizens drive in a particular country, as well as the adoption of electric vehicles and the size of the vehicles. Preferred in every 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 for discovering the air duct forms that make car cabins quieter as well as a machine-learning model, built by the logistics team, that is transported from warehouses to its more than 60 Designed to streamline the shipment of parts to the plants. World.

“It’s a very complex problem – even if you think it might be simple enough,” said Mr. Gur Ari, who said the logistics application is saving the company millions of dollars in shipping costs.

Ford expects that opening up AI development to a wider range of employees could significantly reduce the average time it takes to develop many applications, from months to weeks and even days in some cases.

Philip Kampshoff, a senior McKinsey & Company partner who leads the firm’s Center for Future Mobility, said carmakers could save billions of dollars with the automation and analytics provided by artificial intelligence if they found a way to scale the technology. can understand.

One way to do this is by expanding access to AI development tools and training beyond data scientists and other AI experts, he said, a movement widely known as AI democratization.

Ritu Jyoti, Group Vice President for Artificial Intelligence and Automation Research Worldwide at International Data Corp, said that around 15% to 20% of businesses that have adopted artificial intelligence have initiated AI democratization efforts.

There are challenges in democratizing AI, Ms Jyoti said. For example, the models developed by these AI builders need to be thoroughly validated and tested to avoid unintended consequences. Control and governance are also required to place these AI builders, or citizen data scientists, as they are often called, on data resources and models, on tap and tap to ensure data integrity and appropriate validation metrics such as accuracy, fairness and interpretability. Use it, he said.

Ford said it has solutions. It puts railings around its data resources. It established testing and validation processes 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 Domino’s Data Lab Inc. machine-learning operations platform, which provides a central hub for exploiting compute resources as well as a central hub to deploy and track models at once. An internally developed “feature store” or repository of reusable artificial-intelligence building blocks.