It always is a strange week – that space between the Christmas holidays and the New Year. Romjul – or “Dead Week” – as they call it in Norway (thanks Haje). It’s a time to rest and reflect on the year that was for some – and CES embargoes for others. We’re currently focusing on the latter as we try to keep an eye on the former here at Actuator HQ.
We have spent much of the past few weeks rounding up some of the key trends of the year in this space: delivery, warehouse / fulfillment, and food preparation. We’ve also spoken to a number of key people in the industry, including CMU robotics boss Matthew Johnson-Roberson and Boston Dynamics CEO Rob Playter, OpenRobotics CEO Brian Gerkey and iRobot CEO Colin Angle.
Credit: Paul Marotta / Getty Images for TechCrunch
This week we asked Daniela Rus, Director of MIT CSAIL, to comment on the matter. She really did everything in her answers for this final actuator of 2021 so I’m going to let her get started.
What was the defining robotics / AI / automation trend in 2021? In terms of robotics and automation, the pandemic and the resulting labor shortage have made it clear that robots have a vital role to play in the workplace. Industry has seen increasing adoption of robots for manufacturing and logistics applications, where autonomy can add value, but autonomy on the roads in the form of a robotaxis is still a long way off. Research has made great strides in safer and more powerful robots, with advances in soft robotic bodies and machine-learning robotic brains.
In the field of AI, we have noticed an increased awareness of the challenges of today’s AI solutions. Industry has adopted many applications for deep neural networks that provide tools for improving work in a wide variety of areas. However, it has also been shown that these methods require data availability, ie huge data sets that have to be labeled manually and are not easy to obtain in every area. The quality of this data must be very high, and if the data is skewed or bad, the systems trained on that data will perform just as badly. In addition, these systems are black boxes – there is no way for users of the systems to really “learn” anything based on the inner workings of the AI. There are also robustness problems, as the trained models are often unstable, and we need to understand that the systems do not perform deep reasoning, but mostly perform a flat pattern comparison. The research community is working to address these challenges.
What will 2022 bring for these categories? As we move into our rapidly changing world, I believe that robots and AI can help us develop our human potential as individuals and as a collective. While the last 60 years defined the field of industrial robots and enabled robust machines to perform complex assembly tasks in cramped industrial environments, the next 60 years will introduce robots in human-centered environments to assist humans with physical tasks.
While industrial robots for the past 60 years have been mostly inspired by the human shape, the next steps will be soft robots inspired by the animal kingdom: shape and diversity modeled on our own built environment, with broad potential to mimic our natural state . While industrial robots have been made of hard plastics and metal for the past 60 years, I believe that the next 60 years will bring us machines that are made of materials that are natural to us or through technical processes such as wood, plastic, paper, ice or even food be available .
We’ll also see new ideas for AI, take a more serious look at AI and data protection, and look at AI sustainability. It is important to remember that today’s greatest advances can be traced back to decades of ideas complemented by huge amounts of data and calculations. New technical ideas and funding to support them are needed to ensure progress. In addition, we will better understand the carbon footprint of machine learning and take a more serious look at sustainable AI.
AI innovations can help streamline many of our activities to slow the effects of warming: optimizing the cost of electricity from technology, making transportation more efficient, monitoring and ending deforestation, conserving biodiversity, or ensuring sufficient food. But for this, AI systems have to consume an enormous amount of energy. Current research estimates that training a large deep learning model produces 626,000 pounds of carbon dioxide, the equivalent of the lifetime emissions of five cars. We need to develop simpler models that can drastically reduce the carbon footprint of AI.
The proliferation of robotics, automation, and AI technologies has the potential to make people’s lives easier – but many of the roles these technologies can play will supplant people’s work today. We will also focus on anticipating and responding to the resulting economic inequality.
DEARBORN, MI – SEPTEMBER 27: A Ford Motor Company employee works on a Ford F-150 truck on the assembly line at the Ford Dearborn Truck Plant on September 27, 2018 in Dearborn, Michigan. The Ford Rouge plant celebrates 100 years as America’s longest continuously operating auto plant. The factory produced Eagle Boats during World War I and is currently producing the Ford F-150 pickup truck. (Photo by Bill Pugliano / Getty Images)
It will be impossible to top Professor Rus’ words this week, but I would like to end the year on a couple of points that highlight what happened in the industry over the past year and – maybe – give us some insight into where things are go to lie. I’ll repeat what many people have said on the Actuator pages: It’s been a remarkable year for robotics. After many years exploring space and hearing people predict the end of robotics, the pandemic has resulted in many being put into practice faster than many predicted.
Of course, I don’t think anyone was hoping we got here through a pandemic, but that’s how it works. And, to reiterate what Professor Rus said, we hope that we can address climate change concerns head on before it’s too late – and that we can address the inevitable displacement of human jobs. If we use robots to improve the working conditions of some, we as a society must ensure that we can support those whose jobs are being laid off as a result.
Robotics will tell you about the boring, dirty, and dangerous jobs they want to replace, but there is a far more difficult discussion about the impact change is having on humans. It is true that companies are struggling right now to fill roles – something that automation can and will address. It is also true that most of this technology is still at a collaborative stage that requires human participation. But as technology advances and becomes more powerful, is it inevitable that we will leave large swathes of the population behind because we consider their work unskilled?
When I look at the trends over the past year, these are the main categories I have followed in this area:
- Warehouse / fulfillment
- transport
- Food delivery / preparation
- Agriculture / Agriculture
- home
- Medical / surgical
- Manufacturing
- construction
And in a way, that’s just the beginning. For example we covered Petra a few weeks ago, a drilling company that raised $ 30 million drilling through solid rock for infrastructure projects. The list of fields that robotics will change is long and growing. As we look forward to 2022, let’s think about how we will deal with things like this in the future. I look forward to continuing to discuss these and other issues in Actuator in the future. However, next week we’re going to get a lot more myopic and focus on the robots at CES.
Credit: Meituan’s food delivery drone landed on a pick-up kiosk in Shenzhen / Photo: TechCrunch
Thankfully, the big news about robotics has slowed down a bit in this romjul (I can only use my new word once a year so I make the most of it). Rita has a great story for you about Meituanbringing a drone-based grocery delivery to Shenzhen. The Tencent-backed company delivered 19,000 meals to 8,000 customers during its two-year pilot. As the article notes, relatively loose regulations at the manufacturing center are a big part of what has brought it to Shenzhen.
Speaking of China, the country’s Ministry of Industry and Information Technology has helped develop an ambitious five-year plan The aim is to increase the acceptance of robots. This includes a planned sales increase of 20% per year with a focus on manufacturing.
Happy New Year and congratulations on getting through the last one. Not everyone was so lucky. Here is to share more robot news in 2022.
Credit: Bryce Durbin / TechCrunch
Need a head start on your New Years resolution of subscribing to more free robotics newsletters? I received great news for you.
