It was going to be the factory of the future. Dubbed the “Alien Dreadnought,” Tesla’s new manufacturing facility in Fremont, California, was designed to be fully automated — no humans need apply. If all went well, AI-powered robots would enable the company to achieve a weekly production of 5,000 Model 3 electric cars to keep up with burgeoning demand. But Tesla fell far short of that mark, manufacturing just 2,000 vehicles a week. The problem, as the company painfully discovered, was that full automation wasn’t everything it was cracked up to be. According to CEO Elon Musk, the sophisticated robots actually slowed down production instead of speeding it up.
Tesla’s solution was to shut down production to address the bottlenecks and then to erect a large temporary structure — essentially a tent — for additional capacity. The company has also hired hundreds of workers to revamp production processes, train (and retrain) the robots, and swap them out when needed, among other tasks. As Musk himself tweeted last April, “Yes, excessive automation at Tesla was a mistake. To be precise, my mistake. Humans are underrated.”
Tesla is not the only company to learn the pitfalls of excessive automation. In our global study of more than 1,000 companies at the forefront of implementing AI systems, we have found that the greatest performance gains are achieved not when machines are used to replace employees, but when they are deployed to work alongside them. In such collaborative relationships, people help machines become better, and machines enable people to achieve step-level increases in performance.
Adding Humans to the Mix
For Tesla, adding more human labor to the mix means extending traditional jobs with additional responsibilities that would help ensure the smooth and efficient operation of the Alien Dreadnought. So, for instance, an equipment maintenance supervisor must be able to do more than just supervise hourly technicians and manage the repair of equipment. The worker must also possess robotics and controls engineering skills, according to our analysis of Tesla’s recent recruiting efforts. Similarly, equipment maintenance technicians need more than just the know-how to diagnose and troubleshoot industrial equipment. They must also be able to use a variety of analytics, such as thermography and vibration analysis, to proactively determine when certain maintenance procedures should be performed on machinery before a breakdown occurs.
Sponsored by SAS
How companies are using artificial intelligence in their business operations.
And it’s not just traditional jobs that are being extended to encompass new tasks. Our analysis has uncovered that entirely new categories of jobs are being created. Just as the internet revolution ushered in completely novel jobs — for example, web designer and search-engine optimization engineer — so will the new era of AI. Telsa, for instance, is recruiting robot engineers, computer vision scientists, deep learning scientists, and machine learning systems engineers. And the company has also posted job listings for more-esoteric AI specialties such as a battery algorithms engineer and a sensor-fusion object tracking and prediction engineer. For the former position, the requirements go beyond knowledge of lithium-ion cells (cell capacity, impedance, energy, and so on) to include expertise to develop algorithms for state-of-the-art feedback control and estimation. Moreover, it’s not just technology-related jobs that are being reimagined with AI. In fact, as Tesla and other companies have discovered, AI technologies are having a profound impact throughout the enterprise, from sales and marketing, to R&D, to back-office functions like accounting and finance. As just one example, Tesla deploys an AI system to process its customer data, including information from an online forum, in order to identify common problems with the company’s vehicles.
Some Training Required
Obviously, finding the right individuals to fill roles like “battery algorithms engineer” is not an easy task, especially given the severe shortage of AI expertise, which has pushed some annual salaries well above $300,000. As such, many companies are trying to grow the talent they need in-house. Yet in our global study, we found that although executives have realized that their reskilling programs will require a bigger and different set of activities than in the past, nearly three-quarters of the 1,500 global companies we surveyed said they have struggled with how to proceed.
The solution will require significant new investments in reskilling — especially given that only about 3% of companies are planning along these lines — and may call for collaboration with outside partners as well as government agencies. Consider Adidas’s “Speedfactory,” an advanced manufacturing plant that recently started production outside Atlanta. To open the 74,000-square-foot robotic plant, which will enable manufacturing flexibility for making sneakers designed specifically for local consumers, Adidas worked closely with local authorities in Georgia and with German-based partner OECHSLER Motion. Currently, the facility employs about 150 people in numerous jobs that are highly technical: planners, engineers, stitchers, and technicians. As the factory was being built, OECHSLER staff worked from a startup hub that was run as a partnership between Chattahoochee Technical College, the Cherokee Office of Economic Development, and the Woodstock Office of Economic Development. Other incentives included a state tax credit of $3,500 per job created, as well as assistance from Georgia Quick Start, a state program that provides training support. In addition, Adidas flew employees to Germany for training, to work with the specialized AI-based robotic machinery.
As the amount of employee training increases, some companies have begun to develop their own certification programs to help employees acquire the knowledge and expertise they’ll need. Take, for example, GE Global Research, which has set up online programs to teach machine learning and other specific skills. Several hundred employees have already completed the company’s certification program for data analytics, which have enabled people to assume new roles.
Back at Tesla, Model 3 workers receive more training than other production staff, and this includes classroom training in both manufacturing essentials and manufacturing fundamentals. Tesla has also been launching new technician training programs that, for example, help people make the transition from working on internal combustion engines to electric vehicles. And the company has partnered with colleges to provide students with the education they’ll need for a career in the electric-vehicle industry.
As much as Tesla has embraced automation and AI, the company’s success will ultimately depend on humans. To meet burgeoning demand for the Model 3, Musk has expressed his desire to eventually run three shifts of manufacturing a day, essentially keeping the assembly line in nonstop operation. To accomplish that, the plan is to hire about 400 employees a week, resulting in considerable demands for onboard training to accommodate that influx. Meeting that challenge of employee training will be crucial to attain the necessary economies of scale, given the Model 3’s relatively low entry price point, starting at $35,000. According to one analysis, the car has the potential to achieve a 30% margin, which would be unprecedented for a battery-powered vehicle. Yet even as the company finally achieved the targeted production of 5,000 vehicles in the last week of June, whether it can maintain and accelerate that aggressive pace remains to be seen. Ironically, even in the factory of the future, humans may be needed now more than ever.