Humanoid robots have generated quite a buzz recently, for two related reasons:
- The rapid developments in AI, including generative AI. These technologies enable machines to “learn” tasks faster than ever before, interpret situations and react accordingly. What is more, AI could allow us to give machines instructions via our most human way of communication—natural language. Add to this the fact that key components enabling humanoid robots—batteries, perception and motion sensors, motors, actuators—have become cheaper and lighter in recent years.
- Humanoid robots have emerged as a critical new frontier for the AI and tech industry. Consider the high-profile investment of $675 million in Figure AI for humanoids from big tech giants Nvidia, Amazon, Microsoft and Open AI. Or Tesla showing off new footage of its humanoid bots walking in lockstep. (For full disclosure, the company I work for—Accenture —has also made a strategic investment in this space.)
Does that mean a solution has arrived for manufacturers suffering from labor shortages and looking to automate operations even further?
Still in pilot mode
We have seen first deployments of humanoid robots in logistics and auto manufacturing because of their mobility and manipulation, which allows them to work in synchronicity, performing tasks that are dirty, dull or dangerous for humans.
Take the humanoid robot from Agility Robotics, called Digit, which is in use at the Amazon warehouse near Seattle, lifting and transporting bins that weigh up to 35 pounds, and at the Spanx factory hauling totes onto a conveyor. Companies like BMW and Mercedez are exploring the use of humanoid robots in automotive manufacturing environments for the vehicle assembly process.
I would be lying if I said such advancements aren’t exciting. However, humanoids’ scalability and commercialization, particularly in manufacturing, remain to be seen.
For starters, humanoid robots are far from a bargain, at least at this point, ranging from $30,000 to $150,000 per robot—a stretch for those who want these robots in high quantities. Of course, with a growing installed base, we can expect prices to decline. Another option is the robot-as-a-service model. For example, Agility offers Digit for $30 an hour.
Also consider that, while AI is making humanoids increasingly capable of taking on tasks that require higher-level cognitive abilities, this doesn’t mean they will become the better industrial robots. In fact, it’s very unlikely. Industrial robotics is a highly advanced discipline, which has continuously improved manufacturing over the past 50 years.
Future roles
The question to ask is whether and where humanoid robots can add something to the manufacturing process. There are use cases where skilled labor is in short supply but can’t be automated with traditional robots.
- I could see humanoids playing a role in complex low-volume production, which requires frequent task and product changes as well as dexterity. The promise of humanoid robots is their (eventual) ability to be repurposed for many different activities without re-programming.
- Humanoid robots are best suited for brownfield environments designed for humans. Here, I’m referring to existing heavily automated factories and warehouses where further automation is prohibited by the limitations of current industrial robots and setting up greenfield facilities is cost-prohibitive.
Preparing for humanoids
While the future of humanoid robots in manufacturing is yet to be seen, industrial companies should be considering four things today:
- Do we understand humanoids’ potential? Building partnerships with emerging humanoid vendors and tapping into R&D initiatives with universities can help you learn what humanoids are already capable of and get a sense of what may be possible in the foreseeable future. Ambitious organizations can consider creating an internal team that starts testing and piloting humanoid robots in their facilities.
- How does the potential of humanoids map to the (future) needs of your organization? Include them in your automation strategy and focus on where further automation has so far been elusive. Factor humanoids into your factory or warehouse modernization and retrofitting plans by focusing on where they could provide benefit. Identify which of your processes require high degrees of flexibility and adaptability and map them against your outlook of available labor to understand where humanoid robots could be an alternative.
- Which talent do we need? Humanoid robots won’t be self-sufficient any time soon. If an organization wants to pilot or even deploy them, it needs engineers with the skills to integrate humanoids into its processes and monitor, maintain and repair them.
- Is our technology humanoid-ready? These robots will be a piece of the automation puzzle, working alongside industrial robots, cobots and autonomous mobile robots. This means organizations need a platform to orchestrate their robot operations. Do you already have it and can humanoids be integrated? Also consider what needs to be done to prepare the software stack to enable humanoids in your operations.
No doubt the discussion around humanoid robots will continue. Labor shortage is becoming an issue in many countries and industries, which means humanoids could become essential to reinventing work—supporting and complementing workers, not replacing them.
Brian R. May is North America lead, Industrial, at Accenture.