If you think that volunteer efforts aren’t able to make a dent in the problems, here’s the most recent (4/20/20) production tally for the group Open Source COVID-19 Medical Supplies, who speak regularly on this web series. One volunteer group has tallied efforts by volunteers across 45 countries who have so far produced 2,315,559 pieces of PPE. And that’s not counting the #distillmyheart masks. Here’s Rachel’s recent interview on KTVU. Those masks aren’t going to make themselves, people!

We also heard from Robin Murphy, Expert in Rescue Robotics & Raytheon Professor at Texas A&M University, who updated her slides on the way in which robots are being used in COVID-19 response. You can find more information on Robotics for Infectious Diseases, an organization formed in response to the Ebola outbreak and chaired by Dr Murphy. There is also a series of interviews answering any questions a roboticist might have about deploying robots with public health, public safety and emergency managers.

Next we heard from Missy Cummings, Expert in Robotics Safety & Professor at Duke University. “I’ve been doing robotics, certification testing and certification for almost 10 years now. I started out in drones. And then kind of did a segue over into driverless cars and I also work on medical systems. So I work in this field of safety critical systems, where the operation of the robot in terms of the drone or the car or the medical robot, it can actually do damage to people if not designed correctly. Here’s a link to a paper that I’ve written for AI Magazine that’s really looking at the maturity of driverless cars.“

“I spent a ridiculous amount of time on Capitol Hill trying to to be a middle ground between, yes, these are good technologies. We want to do research and investment and keep keep building a capacity. But no, we’re not ready to have widespread deployment yet. And I don’t care what Elon Musk says you’re not getting full self driving anytime soon.”

“Any reasoning system has to go through four levels of reasoning, you start at the basics, what we call skill based reasoning, then you go up to rule knowledge and expert based reasoning. And so where do we see that in cars? When you learn to drive you had to learn skill based reasoning, which was learning for example, how to track light lines on the road. Once you did that, maybe 20 minutes to learn that then you never actually have a problem with that again.”

“So once you have the cognitive capacity that you’ve learned skills, then you have enough spare mental capacity to think about rule based reasoning. And that’s when you start to understand, Okay, I see this octagon in front of me, it is a stop sign it’s read, I know that what it means there’s a set of procedures that go along with stopping and I’m going to follow those when I see it. Then once you have the rules of the environment that you’ve learned, then you have the spare capacity to start thinking about knowledge base reasoning, the big jump between rule and knowledge base reasoning is the ability to judge under uncertainty. So this is where you start to see the uncertainty arrow growing. So when you go up to knowledge base reasoning, you are starting to have to make guesses about the world with imperfect information. “

“So I like to show this picture of a stop sign partially obscured by vegetation. There are many many, many driverless car system computer vision systems right now that cannot figure this out that if they see some level of partially obscured stop sign, they just cannot see because they don’t see the way that we see. They don’t see the complete picture. And so they don’t judge that there’s a stop sign there. And you might have seen the recent case of the Tesla being tricked by a partially modified 35 mile per hour sign with a little bit of tape to make it see 85 miles per hour. It’s a really good illustration of just how brittle machine learning deep learning is when it comes to perceptual based reasoning. And then we get to the highest level of reasoning where you really have to make judgments under maximum uncertainty. “

“I love this illustration of this stop sign with these four different arrows. You cannot do this, you cannot turn left, you cannot go right, you cannot go straight and you cannot go back. So I’d be curious, I’d like to see what any driverless car would do in this situation. Because what do you do in the situation? You have to break one of these rules, you have to make a judgment, you have to figure out what is the best possible way to get yourself out of the system. And it means that you’re going to have to break rules inside the system. So the expert base reasoning knowledge base reasoning there is what we call top down reasoning, it’s you taking experience judgment in the world that you’ve had. As you’ve gotten older in life and had more experiences, you bring that to bear to make better guesses about what you should do in the future.”

“Bottom up reasoning is, is essentially what is happening in machine learning. You’re taking all the bits and bytes of information from the world, and then processing that to then make some kind of action. So right now Computers are really good at skill based reasoning, some rule based reasoning, humans are really the best at knowledge and expert based reasoning. And this is something we call functional allocation. But the problem is there’s a big break between rule and knowledge. Driverless cars cannot do this right now, until we can make that jump into knowledge and expert base reasoning. What are we going to do have to do?” [Missy Cummings, Robot Safety Expert]

Michael Sayre, CEO Cognicept said “I’m working with a company called Cognizant as CEO. We’re essentially solving a lot of the problems that Missy highlighted, which is that, you know, when we look at autonomy, and moving robots into the real world, there are a lot of complexities about the real world that cause what we call edge case failure in these systems. And so what we’ve built is essentially a system that allows a confused robot to dial out for human operator.”

“Human in the loop is not a new idea. This is something that self driving cars have used. What we’ve built is essentially a system and a service that allows this confused robot to dialogue for help on real time basis. We essentially listen for intervention requests from robots. So that can be an error code, or, you know, some kind of failure of the system timeouts, whatever it is really, we can listen for that event. And then we cause a ticket to be registered in our system, which our operators will then see, that connects them to the robot, they can kind of get a sense of what’s happening in the robots environment, they can get sensor information, populated in a 3d canvas, we can see videos and so forth, that allow the operator to make judgments on the robots behalf. “

“Self driving vehicles is probably not the best example for us. But maybe you would be able to use our system in something like a last mile delivery vehicle, which will face a lot of the same problems. Maybe the robots uncertain about whether it can cross the road. We can have a look at the video feed from that robot, understand what the traffic signals are saying, or what the environment looks like. And then give the robotic command to essentially help it with getting past the scenario that caused that Case failure. So we see this as sort of a way to help get robots into more useful service.”

“You know, right now, even at 1% failure rate for a lot of these applications can be a deal breaker. You know, we, especially for self driving cars, as everybody mentioned, you know, the cost of failure is really high. But even in other sort of less critical cases, like in building delivery, you see, you know, if something is spinning around in a circle or not performing its job, it causes people to lose confidence in the system stopped using it. And it’s also you know, during the time that it’s confused, not performing its function. So we essentially built this system as a way to bring robots into a broader range of applications and improve the sort of uptime of the system so that it doesn’t get into these positions where it’s stuck during its operation.”

“Similarly, we have robots that get lost in spaces that are widely variable. So you know, a warehouse that has boxes or pallets that move in and out of the space very frequently. That’s going to confuse the robot because its internal map is static in most cases. And when you have a very large change, the robots going to be confused about its location, and then not be able to proceed with its its normal operation. That’s something that we can help with we essentially will be able to look at the robot’s environment and understand where it is in its space and then update its location. Again, you know, we look at different types of obstacles. You know, a plastic bag is not really an obstacle, we can, you can run through that. But on a LIDAR, it shows up the same way as a pile of bricks.”

“So by having a human in the loop element, we are able to sort of handle these edge case failures and get robots to perform functions that they wouldn’t otherwise be able to perform and be useful in applications that were maybe too challenging for full autonomy. I think a lot of it has to do with sort of how dangerous is the robot in question. So, you know, for a self driving vehicle, very dangerous, right, we’ve got a half ton of steel, you know, moving at, you know, relatively rapid speeds. This is a dangerous system. “

“On the other hand, in building delivery robots, we’re doing some work in quarantine zones, making deliveries in buildings that allow social distancing to be maintained. We can put needed supplies inside of this delivery robot and send it in a building to the delivery room. So worst case scenario, we might bump into somebody. It’s just inconvenient and might sort of ruin the either the economics of the usefulness of the robot. That would be a good case for these less critical systems. So things like in-building delivery, material handling and logistics spaces. Maybe like a picking arm like a robot arm pulling things out of a box and putting it into a shipping container, or into another robot for in-building delivery.”

“While we try to get as fast as we can, you’re still talking about 30 seconds, maybe before you can really respond to the problem in a meaningful way. So, you know, 30 seconds is an eternity for a self driving vehicle, whereas for an in-building delivery robot, it’s not a big deal. So I think you know, the answer to that it’s pretty application dependent and also system dependent, you know, how dangerous is the system inherently?” [Michael Sayre, CEO of Cognicept]

Rex St John, ARM IOT Ecosystem Evangelist presented an unusual COVID-19 response topic. “This isn’t quite a robotics topic. But a few weeks ago, I began working on a project called Rosetta@home. So if you’re not familiar, there’s a lot of researchers that are studying protein folding, and other aspects of biological research. And they don’t have the funding to pay for supercomputer time. So what they do is they they distribute the research workloads to volunteer networks all around the world through this program called boink. So there’s a lot of these programs, there’s SETI@home, and Rosetta@home and Fold@home. And there’s all these people that volunteer their extra compute cycles by downloading this client. And then researchers upload work, jobs to the cloud, and then those jobs are distributed to these home computers.”

“So because I work at arm, we realize that Fold@home and Rosetta@home. are two projects which are used specifically to study protein folding. They did not have arm 64 bit clients available, which means you can’t run them on a Raspberry Pi four, you can’t run them on some of the newer arm hardware. So there are a lot of people in the community that were wanting to help out with Fold@home and Rosetta@home, which are now being used extensively by researchers specifically to study COVID-19. So we put together this community project. And it came together very, very quickly. Because once everybody learned about this opportunity, they jumped on board very quickly. So what happened was these guys from Neocortex, it’s a startup out of San Jose. They jumped on this and their CTO ported all the key libraries from Rosetta@home to arm 64 and then within a couple weeks a week or two actually, we’re now up to 793 arm 64 bit devices that are supporting researchers studying COVID-19 so anybody that wants to help out if you’ve got a Raspberry Pi four or an arm 64 bit device you can install Rosetta@home on your Raspberry Pi four and begin crunching on proteins to help researchers fight back COVID-19. https://www.neocortix.com/coronavirus

“You can see this is the spike protein right there of COVID-19. COVID-19 uses the spike protein to sort of latch on to the the receptors of human cells and that’s how it kind of invades your body. So they’re doing a lot of work to understand the structure and behavior that spike protein on Rosetta and Fold@home.” [Rex St John ARM IoT Evangelist]

Ken Goldberg, Director of CITRIS People and Robots Initiative said “I do have one thought that I’d like to share that occurred to me this week, which is that I wonder if we’re shifting from what used to be called the ‘High Tech High Touch’ concept from John Naisbitt. He wrote ‘Mega Trends’ about how we were moving as we got toward more high tech, we’d also just as much crave that touch. And I wonder if we’re moving toward a low touch future where we actually will see new value in things that are don’t involve touch.” “It’s been so interesting for me to be you know, to be in the house. I’ve gotten a whole new appreciation for things like washing machines and even vacuum cleaners. They’re incredible these mechanisms that help us do things, that rather than us reaching down and touching everything they basically do it for us.”

“I’ve been thinking about you before this pandemic. There are a lot of things out there like robot vending machines that I was a little skeptical about. And I thought, well, I don’t really see what’s the big advantage, given a choice I’d rather have a human making a hamburger or coffee. But now I’m starting to really think that equation has changed. And I wonder if that’s going to change permanently. In other words, are we actually going to see this a real trend toward things like these robot coffee making baristas and robot burgers like Creator, the company in San Francisco. or Miso robotics is developing fast food making robots. I think it’s time to really reevaluate those trends because I think there is going to be an actual visceral appeal for this kind of low touch future.” [Ken Goldberg CITRIS People and Robots]

There’ll be more next week on Tuesday April 28 so sign up for COVID-19, robots and us with guest speakers focusing on regulations, risks and opportunities:

• Chelsey Colbert, Policy Council at The Future of Privacy Forum
• Michael Froomkin, Laurie Silvers and Mitchell Rubenstein Distinguished Professor of Law
• Ryan Calo, Lane Powell & D. Wayne Gittinger Endowed Professorship of Law
• Sue Keay, Research Director at CSIRO Data 61 and Aust National COVID response
• Robin Murphy, Rescue Robotics Expert and Raytheon Professor at Texas A&M University
• Thomas Low, Director of Robotics SRI International
• Ken Goldberg, Director of CITRIS People and Robots Initiative
• Andra Keay, Director of Silicon Valley Robotics and founder of Women in Robotics

“Grocery workers are risking their safety, often for poverty-level wages, so the rest of us can shelter in place,” said John Logan, director of labor and employment studies at San Francisco State University. “The only way the rest of us are able to stay home is because they’re willing to go to work.” [Washington Post April 12 2020]

In our April 7th edition of “COVID-19, robots and us”, we heard from Linda Pouliot, CEO and founder of Dishcraft Robotics and Vipin Jain, CEO and founder of Blendid. Both provided us with insights into robotics in the food industry and the difficulties and joys of adapting robots to COVID-19 work.

Dishcraft Robotics is Linda Pouliot’s fourth startup, and second robotics startup. She previously cofounded Neato Robotics, the number two player in automated vacuuming. Dishcraft Robotics is a clean dish delivery service for cafeterias, or large kitchens. On a daily basis they deliver customized clean dishes to the customer, they collect the dirty dishes, and return them to a central hub where custom robots wash them.

“We initially thought we would develop the robot and install it directly into dish rooms. And then we found that it was a really high hurdle for an operation to have a big capital expense, and that we could solve their problem in a really frictionless way, by simply doing the delivery because there’s no upfront cost for them. There is no risk for them to try it. And in fact, we give a free couple week trial and we have found that once you start with dish craft, you immediately convert because the service solves all their labor problems, it also is more sustainable. And in today’s environment, it’s very, very sanitary.”

“We give them carts full of clean dishes, they are our dishes that are proprietary. The way our robots work is using magnetics. And so it requires using our specialized wares. And so we just bring in flexible, clean dishes and we give them a collection system where they can drop off. It’s very organized, it’s ADA compliant. It’s pretty space saving compared to where people normally store their dishes. So it’s worked really well for us. And that means for a customer, there’s no construction costs. There’s no months of planning. You know, it’s just been pretty delightful for both us and them.”

“Currently we service essential businesses and we have started to offer our services to help in this time of need to hospitals. Dish rooms are very small and cramped spaces and we realized that workers in them were unlikely to be able to be socially distance. Also today, the hospitals are overloaded. Some have twice the volume of what they normally expect a dish room to have to process. And so this was a great opportunity for them to use robots and use our service and keep their existing staff very safe.” [Linda Pouliot April 7 2020]

Vipin Jain is the CEO and founder of Blendid, a food robotic kiosk startup. After about 5 years of R&D Blendid opened their first kiosks just over a year ago in March 2019 at the University of San Francisco, followed by Sonoma State University, and while they were initially doubtful that people were ready to have their food or drink made by robots, they’ve seen great interest.

“We were told that automating food was too hard. And people won’t like it. People don’t want robots to make the food, people said. But people have different tastes and and standardized food doesn’t cut it anymore. So you need a better solution so that I can get what I like to eat anytime of the day wherever I am. And that can only be done using a lot of data, AI and robotics.”

“We were just getting ready to start deploying into retail, corporate and hospital environments when the COVID hit. So I wish we were a little bit ahead further ahead in terms of deployment because as you can imagine, food is an essential service. We all want food, we all want access to food, but in this environment, we will be fearful about the safety of people working on food preparation, because we don’t know how the food was handled.” [Vipin Jain April 7 2020]

Rich Stump is one of the cofounders of Studio Fathom for prototyping, product development and production services. Fathom specializes in 3d printing additive manufacturing, dealing with multiple materials and leveraging traditional manufacturing alongside, allowing them to solve some interesting engineering challenges or product development challenges. Fathom has been very active in COVID-19 response.

“There’s three main initiatives that we have going right now with COVID-19. And some of them are, are potentially revenue generating, but most of them are just just trying to help us with our expertise and capabilities. The first is obviously with PPE challenges we have. We have a facility in Asia and southern China and so we have a vast amount of supply chain resources. So, we called in our colleagues over there and we found a number of factories that had FDA approved PBE supplies that that were overrun from the flattening of the curve in Asia.”

“The news has been talking about investment into 3M and Honeywell ramping up production in the US, which is absolutely great. The problem is, it’s just not going to get done in time, right? Anytime you buy 150 thermoforming machines and you try to ramp up production to make n95 masks, it’s going to take weeks or potentially months to get any volume. So what we were trying to do is match hospitals and senior homes and folks that needed PPE with supply chain resources that we had in Asia.”

“We’ve had a bunch of roadblocks to the point where we’ve tried to reach out to the state government officials to try to remove some of the roadblocks with customs and tariffs and freight constraints. And so that’s been a an interesting challenge. But we’ve been able to connect, I think over 1.5 million and counting supplies into the US from Asia. It’s not our everyday business, but we have resources, so we tried to help there.

“The second initiative was around the test kits. Obviously there’s been a shortage of test kits. And it turns out, interestingly enough, one of the big shortage items is the nasopharyngeal swab that goes back into your nasal cavity. The main manufacturer was in Italy, and obviously with everything happened in Italy, that that shut down a lot of the supply. There’s been a community of about 50 folks that have come out of the 3d printing community to see if we can 3d print these swabs. And we’ve made a lot of progress.”

There’s now four manufacturers approved to 3d print nasopharyngeal swabs. Next week, we’ll get into production of these in at least hundreds of thousands, if not millions, in order to support the short term need. Long term, it probably doesn’t make sense to 3d print them given the cost base. But using 3d printing technology, we can design swabs that could potentially perform better than the traditional swab. So that’s been a fun and challenging project at the same time.”

“And the third initiative is around the ventilators. You have folks like Lawrence Livermore developing a ventilator with spare parts that are available today. And you have JPL, the NASA team, trying to develop their own ventilator. Then you have the large automotive manufacturers who have partnered up with various existing ventilator manufacturers and are trying to ramp up production. By today I think more than 270 people have come to us with either a ventilator design, a mask design, or some type of shield design, just over the last 10 days. There’s just a tremendous amount of activity. And it’s great because everyone wants to help, but at the same time, it presents a lot of challenges because so much effort going in so many different directions, that you you worry that, you know, all these efforts won’t end up being impactful by the time we need the supplies.” [Rich Stump April 7 2020]

Mark Martin is the Director of Industry and Workforce Development for California Community Colleges and founder of the Bay Area Manufacturers Discussion Forum. He’s very connected with manufacturers in the region and has a lot of insights into what’s going on in the manufacturing sector.

“Many of the manufacturing companies are being hurt and having to furlough people. There’s about 8000 in the Bay Area and we’re trying to help them repurpose. In some things, it’s relatively easy, maybe you have an injection mold facility, and you can injection mold these face shields, or some of these parts of the face shield. You have to make the molds but it’s still within your core business. Others are contract manufacturers that could theoretically assemble ventilators because they’re already doing medical products. Ventilators are a complicated product, but not necessarily more complicated than some of these people already doing.”

“But for others, you might have to get specialized machines, which can take a long time. And then you need to have expertise around. If you haven’t done medical products before, do you know how to handle the equipment and ensure QC? And the FDA approves factories, not just the designs. I created a list of Bay Area manufacturers and I got almost 200 responses with things they think they can manufacture and I’ve supplied that list to the state government.”

“And community colleges have makerspaces or fab labs. So where I’m located at Laney College in Oakland, we started working on face shields a week and a half ago, because that was relatively simple for us to be able to do. And we printed up 500 face shields to Highland hospital. It took us a day to do it with six people. And I asked the hospital what their daily use was for face shields and they said 600 a day. So I was like, all right, that was a day’s worth for this one hospital.”

“And then we actually took a design that was done by Stratasys for 3D printing and modified it a little bit. And in like two days, we got a molder to set up injection molding, although it’s going to take a couple of weeks to finish. And honestly, I have no idea if we’re going to need in a couple weeks. Because I have no idea what the demand is.”

“Apple’s bringing in over a million a week and others are doing this hundreds or even thousands a time in basements and little shops. But we don’t know if the demand is 5 million a week or 50 million. “

“So that that’s the thing the manufacturers all trying to figure out. I kept wondering why the government didn’t take a little bit of control. Even just financial incentives to say we will backstop you if you supply PPE, so that if the demand falls out, you don’t lose money.” [Mark Martin April 7 2020]

Ken Goldberg is the William S. Floyd Jr Distinguished Chair in Engineering at UC Berkeley, Director of CITRIS People and Robots Initiative and a regular guest on COVID-19, robots and us. Today in keeping with the theme of food robotics, Ken explained the reasoning behind one of his latest projects, AlphaGarden.

This is a farm bot that you can buy from a local company, in San Luis Obispo for about $3,000. Our garden includes 14 species of edible herbs and plants and they’re all growing in close proximity. The challenge is that they’re competing for increasingly scarce light and water. So what happens is that this starts to get quite out of control. Right now, the garden has been acting very autonomously because we haven’t been able to get into the greenhouse, and so it’s starting to decay, and certain invasive species like mint have essentially taken over.”

“We’re trying to optimize diversity, that is to maximize the number of different plant types that are growing. But the key idea here is that we’re using AI to simulate the garden, because a grow cycle takes three months. So we have a simulator that can simulate 10,000 times the speed of nature. We have 64 gardens but thousands of parameters to fine tune the system. Then we evaluate each of those gardens to figure out how to tune the parameters for the watering and pruning devices.”

“This is an ongoing project but the end goal is to be able to learn a policy for successful gardening because polyculture is much more labor intensive than monoculture. The reason that I call it an art project is because it’s extremely difficult for AI and robotics. It’s a big challenge. I’m not at all convinced we’re actually going to succeed. We’re really putting AI to the test. But just this week we did a deep learning method based on the data that we collected from our simulator. And in time we hope to learn how to be able to automate some of these very difficult tasks like organic gardening.” [Ken Goldberg April 7 2020]

The moderator of the weekly ‘COVID-19, robots and us’ discussions, Andra Keay, is the Managing Director of Silicon Valley Robotics, supporting innovation and commercialization of robotics technologies. Andra was also the Industry Co-Chair of the Human-Robot Interaction Conference 2020, which is taking place online. In the April 3 Industry Talks Session, Chris Roberts from Cambridge Consultants described the process of working out what sort of robotics or automation made sense for food production.

“Chris described a project with the food service industry to look at using robotics and AI. And of course what he wanted to build was a robot dipping chips and flipping burgers and all the rest, but on evaluation, that was not the best direction to go. In this instance AI optimization of locations was the best option. And this is something I’ve heard from a number of other people who’ve been doing robots for the actual preparation of food itself. For example, Creator, who make those fantastic hamburgers, didn’t have a robot arm, imitating the way humans cook. Instead, they completely re architected the entire process of constructing a burger, so that it could be done mechanically.” [Andra Keay April 7 2020]

There’s much more to hear in our weekly discussion on ‘COVID-19, robots and us’ from April 7 2020 with guest expert speakers: Linda Pouliot, CEO of Dishcraft Robotics Vipin Jain, CEO of Blendid Rich Stump, Principal at Studio Fathom Mark Martin, Industry & Workforce Development California Community Colleges and Bay Area Manufacturing Discussion Forum Ken Goldberg, Director of CITRIS People and Robots Initiative and William S. Lloyd Jr Distinguished Chair in Engineering at UC Berkeley; moderated by Andra Keay, Silicon Valley Robotics.

Robin Murphy, Raytheon Professor at Texas A&M University and founder of the field of Rescue Robotics, was involved in the very first use of robots in a disaster scenario in 9/11. Since then she’s been involved in multiple disaster responses worldwide, including the Ebola outbreak in 2014-2016. During the US Ebola outbreak, the White House Office of Science, Technology and policy, and then later NSF did a series of workshops, and myself and Ken Goldberg, are among those who participated in work with various public health officials in groups such as Doctors Without Borders.

“Some of the lessons learned about infectious diseases in general, and for COVID, in particular, are that there’s really five big categories of uses of robots. Most everybody immediately thinks of clinical applications, or things that are directly related to the health care industry, but the roll of robots is far broader than that. It’s not always about replacing doctors. It’s about how can robots assist in any way they can in this whole large, complex enterprise of a disaster.” [Robin Murphy March 31 2020]

Ross Mead’s company Semio develops software for robot operation that focuses on how people will live, work and play with robots in their everyday lives. “We’re building an operating system and app ecosystem for in home personal robots. And all of our software relies on natural language user interfaces, just speech and body language or other social cues.”

“Right now as it pertains to COVID-19. We are working with a team of content creators from a company called copy to develop conversational content similar to chatbots, or voice voice based skills that’s geared towards informing users about or helping mitigate the spread of COVID-19. We’re also developing socially aware navigation systems for mobile robots, natural human environments. I would love to talk about use cases for social robots, even telepresence robots, as well as the impacts of social isolation in these times.” [Ross Mead March 31 2020]

Wendy Ju studies people’s interaction with autonomous cars, systems or robots. She focuses on things like nonverbal cues, and the things people do to coordinate and collaborate with one another. Her PhD dissertation was on the design of implicit interactions, something a lot of us take for granted or consider static, not dynamic. Through her work on autonomous cars, she’s been exploring the subtle cues that convey critical information.

“if we get these things wrong, or it’s life or death. I think we’re starting to understand that a lot of the things that we think of as interaction are only the top layer of what we’re actually doing all the time with other people. And if we don’t understand those lower layers, it could kill us.”

“So last week, I put together a proposal to study how people are interacting with one another in the city around the social distancing policy. And I agree the name is not perfect, but I think it also gets to the heart of what’s important to do. In this epidemic, there’s a halo effect around our social interactions because we know they’re necessary and good for us. And so people think, well, I shouldn’t go to the grocery stores and go to the hospital. But surely, it can’t be bad to go visit my neighbor or surely it can’t be bad to go see my grandmother, these kind of inclinations will kill us, when taken to scale.”

“When we say social distancing, we’re saying like, yes, school is good, but school is bad in the situation, churches good, but churches are bad in the situation, really getting at the thing that we are so tempted to do that is literally the thing that we’re trying to stop right now. I think that’s why they call it social distancing. And it does definitely have a physical corollary. I’m interested to see afterwards, if those people who were playing basketball and other people who were playing soccer, are those places where people got more sick or not? We don’t actually know all the different mechanisms for transmission for disease. And I think later on, we’ll be able to figure it out.” [Wendy Ju March 31 2020]

Cory Kidd is the CEO and founder of Catalia Health, which uses social robots for medical care management. Catalia Health has done extensive clinical trials prior to commercial roll out and leads the world in understanding robots for medical care.

“The concept of chronic disease management of course is not new, it’s just that the usual model is very human powered. We do it in clinical settings, in the doctor’s office, in the hospital, and we send people out to homes, and a lot of the work is done by calling patients on the phone to check in on them. We replace all of those by putting actually a physical robot in the patient’s home to talk to them.

“So what we’re doing on the AI side is generating conversations for patients on the fly, for whatever condition they’re dealing with, and we build these around specific conditions. The robotic piece of it is really driven by the psychology around why we would rather be all in a room together, as opposed to, gathering around our computers and staring into the the screen at zoom. We intuitively get that physical presence is different. “

“When we’re face to face with someone, they’re more engaged, we create stronger relationships and a number of other things. Research showed that those differences actually carry over into the future. When you put a cute little robot in front of someone that can look them in the eyes while it’s talking to them, we actually get a lot of the effects of face to face interaction. And so we’ve leveraged that to build chronic disease care management programs. Over the last couple of years, we’ve been rolling these out largely in specialty pharmacy, so we work with some of the largest pharma manufacturers in the world, like Pfizer. We’re helping patients across a number of different conditions really keep track of how they’re doing, to stay on therapy and stay out of the hospital using our AI and robotics platform.”

“The current situation around the world is really highlighting the need for more of this kind of technology.” [Cory Kidd March 31 2020]

Ken Goldberg is the director of CITRIS People and Robots Initiative, and the William s. Floyd Jr. Distinguished Chair in Engineering at UC Berkeley. Both Ken and Robin are amongst the authors of recent editorial in Science Robotics “Combating COVID-19 – The role of robotics in manging public health and infectious diseases”.

“There’s this fundamental issue that I’ve been thinking a lot about, which is protecting the health care workers, especially where they’re now having to provide these tests for huge numbers of people. Swabbing is quite an uncomfortable and invasive process. And is there any way that that that we might be able to automate that at some point? I don’t think that’s going to happen anytime soon. But it’s an interesting goal that we could move in that direction.”

“The other is the idea that intubation is an incredibly difficult process and very risky because a lot of droplets vaporizing happens. That’s another area where it would be very helpful if that could be teleoperated. Right now, the state of the art in telemedicine, tele surgery in particular, and these type of procedures is not ready for the situation we’re facing now. We are nowhere near capable of doing that. And so I think this is a really important wake up call to start to develop these technologies.”

Also, in the discussion, Jessica Armstrong who is a mechanical engineer at SuitX and local coordinator for Open Source COVID-19 Medical Supplies gave us updates on local PPE activities and how community grass roots initiatives like OSCMS and Helpful Engineering have been part of catalyzing networks of people to sew masks and gowns, to laser cut face shields and 3D print parts for PPE and medical equipment, and developing new designs for emergency ventilators and respirators, while we’re still waiting for manufacturers and the supply chain to meet the demand.

Perhaps most critically, groups like OSCMS and Helpful Engineering validate and share designs for PPE so that people aren’t wasting time designing their own solutions, nor putting health care workers at risk with badly designed homemade PPE.

Our second weekly discussion about “COVID-19, robots and us” from March 31 is now available online and as a podcast. You can sign up to join the audience for the next episodes here.

Special guests were Robin Murphy, Raytheon Professor at Texas A&M University and founder of the field of Rescue Robotics, Ross Mead, CEO of Semio and VP of AI LA, Wendy Ju, Interaction Design Professor at Cornell, Cory Kidd, CEO of Catalia Health – maker of medical social robots, Ken Goldberg, Director of CITRIS People and Robots Initiative and Jessica Armstrong, mechanical engineer at SuitX and local coordinator for Open Source Covid-19 Medical Supplies. Moderated by Andra Keay, Managing Director of Silicon Valley Robotics, with extra help from Erin Pan, Silicon Valley Robotics, and Beau Ambur from Kickstarter.