The robots are coming (but don’t give up your job yet)

The pandemic accelerated growth in the industrial robotics market as enterprises sought to mitigate COVID-related staff crises with automation. It upturned existing supply chains, exposing weaknesses in the labor market.

As factory wages climbed, manufacturers invested in AI and robotics to improve efficiency. While many industries already used smart machines to automate production, the pandemic accelerated deployment, as borrowing was cheap. The market is expected to reach $31 billion in 2028, up from $14.6 billion in 2020.

Today, robots are appearing across multiple industries. Not just in warehousing and manufacturing, but also in healthcare, conferencing, last-mile logistics and even catering. The nature of a robot – what it looks like, how mobile it is, and how “intelligent” it is built to be – reflects its use. While some are static, others are mobile. Some are dextrous, and others use machine-sensor intelligence for situational awareness. Some robots learn; others perform single, repeated tasks.

Advances in machine learning and engineering mean we’ve hit an inflection point when these intelligent machines enter new sectors. Boston Consulting Group (BCG) believes mobile and stationary professional services robots will become a $170 billion market by 2030, compared to $50 billion for conventional industrial robots/cobots, and $30 billion for automated guided vehicles (AGVs). In many cases, robots will augment human staff, taking on mundane jobs and liberating workers for more rewarding tasks.

How are robots being used today?

In Australia, Alphabet’s Wing subsidiary drones make a delivery every 25-seconds, and Starship Technologies delivers takeaways in Milton Keynes. But last-mile delivery reflects automation across the supply chain. Orange Business, for example, provides 5G network services to support numerous automated systems at the Port of Antwerp, including tugboats and forklift trucks.

Arc and spot welding, materials handling, picking and packing are just some tasks already transacted by industrial robots, particularly in vehicle manufacturing; but as the tech improves, robots are becoming capable of addressing more complex challenges.

Some examples: A Boston Dynamics robot, Stretch, can move 800 boxes each hour, autonomously handling challenging truck loading and unloading tasks. As spatial awareness and augmented reality (AR) enter the mainstream, analysts expect robots will work with humans to handle dangerous missions. In recent months, the use of unmanned aerial vehicles (UAV) in crisis situations has been well reported.

In telepresence, Apple used a Boston Dynamics robot called Spot to help architects oversee the building of its new headquarters in London’s Battersea. Schneider Electric uses a 5G network to connect an AXYN telepresence robot for remote site visits. And while there’s no news yet, the entire world knows Apple, Google, Tesla and most car manufacturers are attempting to build an autonomous self-driven robot car.

But the pandemic accelerated new industry deployments

The need to preserve social distance combined with frequent staff absences due to COVID-19 accelerated the use of robotics in catering. Beijing’s Winter Olympics used a sophisticated robot chef to feed guests and athletes while limiting human contact to protect them.

The automation of culinary life exposes fresh business opportunities: Nala Robotics promises its restaurant-as-service platform will deploy full kitchens in just 24 hours. As labor shortages impact catering, it’s reasonable to expect robots will become as common as any kitchen implement. White Castle (a U.S. burger chain) employs semi-autonomous robot chefs capable of flipping and prepping 300 meat patties a day in some restaurants.

The pandemic also accelerated robot deployment in life sciences where adoption climbed 69%, according to the Robotic Industries Association (RIA). Robots saw extensive use in COVID-19 testing labs and to support nursing and aged care. The latter market may hit $2 trillion value by 2026.

As robots (and the tech that drives them) become smarter, they can be trusted to make bigger decisions. A Johns Hopkins University robot called STAR (Smart Tissue Autonomous Robot) is a good illustration: STAR successfully performed challenging keyhole intestinal surgery on pigs. This form of surgery is difficult to automate, but the developers claim STAR can plan, adapt and execute surgery with minimal intervention.

A Johns Hopkins University robot named STAR has successfully performed challenging keyhole intestinal surgery on pigs. STAR can plan, adapt and execute surgery with minimal intervention.

Opportunity? But what about challenge?

Robots at work pose philosophical and ethical problems. What will the impact of automation be on employment? How can humans verify the decision tree that machines follow? Such questions have inspired standards development from the American National Standards Institute, Association for Advancing Automation and governments, including the EU’s tough ethics guidelines for trustworthy AI.

Robots require reliable, low latency, secure networks. While it is true they must host enough edge intelligence to make decisions out of complex datasets in real time, they also need networks to share data and access supplementary cloud-based intelligence to empower decisions. The inherent capacity to slice 5G bandwidth to create highly secure private networks enables this. It was a solution used across the Port of Antwerp to support its many connected automated systems.

However, as robots become more sophisticated, they may exceed the cognitive capacity of humans working with them. This may generate suspicion and make AI decision making hard to verify. Duke University Professor Mary Cummings believes that tech must reflect the strengths and limitations of humans to avoid this. “The sweet spot is a collaboration between humans and machines,” she said. This partnership should see robots helping to prevent human error, just as humans prevent AI errors. In this model, both robots and humans augment each other. Soft skills work with the machine.

But are there risks to working with robots?

First-generation robots lacked sensors to detect humans, making them dangerous to work nearby. Today, processor, machine learning and sensor improvements enable technologies such as machine vision, presence-sensing and light curtains to improve safety.

This brave new robot world also needs humans for maintenance. Robots may learn to maintain themselves, but humans will oversee complex repairs. BCG estimates maintenance will cost three times the initial investment across the lifecycle of these machines.

Today’s robots are smart, capable of some decisions and can be trained, but they aren’t yet truly autonomous. They still require human trainers, supervisors and colleagues.

Rapid enterprise deployment of “no code” or “low code” programming makes it easier for those humans to teach old robots new tricks. Traditionally, training was complex and time consuming, but these new models enable non-technical staff to teach machines. Other advances, such as image recognition, provide robots with additional data, while edge-computing developments empower faster decisions, even when servers are offline.

Despite these many deployments and technological advances, none of us can give up our jobs yet. BCG expects fully autonomous machines will first appear by 2030 in limited sectors, such as room service or delivery.

Read this CIO’s guide to turning data into value with connected products, and the issues of robot cybersecurity in this CXO survival guide on IoT security. And read about Schneider Electrics’ trials of 5G-enabled robots on its factory floors.