Our second panel of the evening explored the capabilities and application of robots. Our three expert panelists come from varied backgrounds which made for an interesting discussion. Chaired by Phil Westcott, Co-Managing Director at Filament, the panel looked at how robotics are developed and built, what exactly they are doing out in the world of business and, more importantly, how that will affect the rest of us at work.
Phil invited each of our panelists to introduce themselves and share something they have recently worked on in the realms of robotics. Up first was Brooke Roberts-Islam, Co-Director at Brooke Roberts Innovation Agency (BRIA) and Fashion Tech Writer and Lecturer. BRIA creates material-tech collaborations and products, as well as technology-led installations for brands from sectors spanning fashion, technology, healthcare, smart homes and IoT.
To watch the discussion click the video below.
Changing the narrative
Keen to explore the narrative that currently exists around how robots should look, feel and interact with humans, Brooke set out to create an installation that would challenge people’s viewpoints. In September 2017, as part of the London Design Festival, over 60,000 visitors saw BR Innovation Agency’s Slave/Master installation at the V&A in collaboration with Kuka Robitics. The installation explored the "borders" around human/robot interaction, reversing the traditional “fear” portrayed in sci-fi films of robots oppressing and interfering with humans.
The installation used collaborative robots, sometimes referred to as cobots, which were able to work alongside human co-workers (in this case, dancers) and can interact with them safely in a common working environment. Slave/Master combined contemporary dance, cutting-edge robotics and interactive projection graphics to create a collaborative performance between all of these elements, with the audience able to roam freely around the installation space, viewing it from all angles. It’s no longer at the V&A, but you can watch a video below.
“Ultimately Slave/Master was a collaborative project for public engagement,” Brooke explains. “We wanted to ask how robots might change our lives and how will we work alongside them?”
“Sci-fi films often have fantastic storylines but generate fear and misconception around robots. We wanted to challenge and change the narrative. We used new robots that are designed specifically to work alongside humans. We knew we wanted to work with Kuka Robotics because they are one of the leading manufacturers of industrial robots.”
Phil asked what kind of response Brooke received from the installation as she had asked audiences to give feedback. “Oh, we had a fascinating array of responses,” said Brooke “we found that children were much more responsive, wanting to touch the robots and go much closer to them than the adults.” The installation also opened up discussions with companies about using tech. Many wanted to ask how they could use tech and or robots creatively with things like design and painting. “Robots are useful tools but they are also potentially very creative as well,” Brooke believes.
It’s like talking to a robot
Even if you haven’t seen robots dancing lately, there’s a much higher chance that you’ve spoken to one. From the automated voice on self-service checkouts to chatbots on websites that help answer queries and questions, chatbots are becoming more common.
Jacob Tomlinson is Lead Engineer at the Met Office Informatics Lab a specialist team which focuses on solving long term issues within the Met Office by creating prototypes of solutions. A major focus is making the hundreds of terabytes we produce useful for people. Following the successful work done with Microsoft developing a weather and climate bot, the Informatics Lab has been busy creating a more user-friendly, public-facing bot, due appear on Facebook Messenger as a prototype in the very near future.
Jacob leads the technical infrastructure in these projects. As well as designing experiments, evaluating new technologies and architecting big data systems, he also writes, gives talks and shares the lessons learned in the Lab.
“What we do is so much more than the weather,” he says. “People want to know how the weather will actually affect them. They don’t really care if it’s raining if they’ve known ahead of time and brought an umbrella. We’re trying to get the message across provide people with that information, so they don’t have to think too hard.”
“People try and play around with the bot – they treated it like a search engine, modifying and tweaking their questions. We wanted more of a conversational aspect. The bot could maintain more context and conversation than people expected. We need to meet the needs of now and aim for the needs of later.”
Phil wanted to know how Jacob and his team know what kind of things the chatbot would need to be able to understand in order to ‘chat’ with someone and provide accurate information.
“It’s all about testing and prototyping,” answered Jacob. “We got good engagement from chit chat and small talk from the chatbot as well as answering questions search queries. The interesting (and problematic) thing is that people tend to talk to a bot like they would a search engine, with basic keywords instead of conversationally, we want to encourage them to ask questions as they would of another human.”
Another important element to the bot is getting the language right: “We’ve been researching different words and what they might mean to people. For example, if they use ‘probably’ or ‘maybe’ what does that mean in terms of likelihood, 70%? 50%?” says Jacob. A graph of a poll that ranked all possible words pertaining to the certainty of particular type of weather was used. “We then feed that back into the data we share with people—it’s a better way of communicating and it’s more accurate. People tend to care less about the weather and more about how it’s going to affect them. They want to know what decisions they should make.”
“A lot of what we do when we are building a bot is about choice architecture—how you produce a serious of options in text. You only have a small number of characters to use,” Jacob explains. “The team I are always looking for new things on the horizon and how we can use them to improve our organisation. We build a prototype, try it out and learn a load of lessons from that before moving on. Then we might bring in and agency to help us build at scale.”
A great example of this is the team chatbot the lab us to test what we should or shouldn’t add to the weather bot: “We had a discussion on whether or not to add a function to our team bot that told us when the bicycles to go between our buildings were free for us to use. It was going to end up being an over complicated process and wasn’t necessary. Having a place where we can throw out theories like that is really handy.
Rory McElearney, is the Conversational Interfaces Lead at Filament. A front-end engineer and chatbot architect, Rory designs and develops for chat interfaces leveraging cutting-edge Natural Language Processing and conversation management techniques. He has written on conversation UX for D/sruption magazine.
Rory’s role sees him looking at how AI is actually being applied to existing problems in an attempt to solve them. He develops the language the bots use to help people interact with them in a smooth process rather than making life difficult. He writes about his real-life experiences with conversation UX and machine learning bias in a highly entertaining way.
Phil wanted to know what sort of skills make good conversational UXs who should we be hiring for roles involving it? Rory explained that it’s important to have a balance of tech and creative skills: “The UX and the copy you write is so linked to the technology you kind of need a bit of everything, you need par dev., part writer—you need a full stack of skills.” Jacob agreed “you need someone who can communicate effectively. It’s a human problem that needs to be solved with tech. The chatbot falls under public research and digital, so when we hire people to work on it the default would seem to be a graphic designer—but it’s also all about the copy. Peel away all the skills to creative thinking and language understanding.
Rory is excited for the future of AI “we’re working on the first AI/AR mashup for the London Aquarium to augment the experience of going around the aquarium. You can talk to the ‘buddy’ which gives you extra information and facts about exhibits and the various sea creatures on show. It gives us the chance to glimpse into the future. I love the idea of taking an experience that’s happening already and weaving in a tech solution, instead of clumsily putting it in.”
However, Rory warns “we have to be aware of over engineering things. AI should add to the experience but not be added in unnecessarily or just for the sake of it.”
From a business point of view it’s clear that the greater need is for bots that can increase human capability, but you do also have to take into consideration how people might respond to how the robot looks. There’s a slightly uncanny nature to a robot that looks almost human and Phil wanted to know more about responses and reactions the panel have seen when people interact with the bots and AI they have developed.
Brooke began: “it’s interesting how people respond to robots and cobots. We had a stronger response to the robot, the cobot was simply described as ‘cute’ or ‘playful’. Perhaps adults feel more that robots could usurp them or take over—children don’t have that learned behaviour. I think that creative people feel less afraid perhaps because they are used to working around something.”
Rory agreed with Brooke on the generational divide: “Adults were very impressed but children who have been exposed to tech from the beginning—kids now expect touch screens and AR— were less bothered by them. There’s a greater familiarity with tech for younger people now.”
“People love to give feedback,” said Brooke continued. The people we asked to give feedback through a bot weren’t interested in the questions, they wanted to make their point and just say what they wanted to say whether it answered the question or not.”
Rory told the audience “people think that it’s somehow anonymous and private, when in fact a human will be looking at the data and conversations at some point. People made, shall we say, some fruity comments to one of the bots we created!”
“People play around with them and try to get them to say silly things or try to trick them,” added Jacob.
Phil posed “should we give them (robots) a human like character to help people respond to them?”
Rory answered “we might include small talk and give the bot a little more personality, but we would never make one that proports to be a human unless it was core to the mission of the bot—like one that is used to provide therapy. That’s not a target on the wall for us.”
Jacob added “we find that people are more likely to accept mistakes in language or even some information if it comes from a bot. It makes it more forgivable—especially if the bot has character. If it was a human, it would be a different response.”
“Some feedback we had was that people were more comfortable with less human like robots. When androids are mimicking human behaviours, people are extremely uncomfortable. It’s the uncanny nature that makes them nervous,” Brooke noted.
Phil agreed “we’re all sort of fascinated and terrified when there’s a robot that looks like a person!”
Of course, we cannot know the future, but it’s clear that AI, AR, robotics and tech are already becoming a bigger part of all our lives. Businesses are employing certain aspects to influence sales and engage with clients and customers in a new way. Phil cited the Charlotte Tilbury magic mirror as a way that AR is becoming more of a commercial tool used to sell products. “There are so many applications of AI and AR and I think businesses will become more and more used to employing these techniques as another way to reach people.”
Jacob agreed “most people have weather apps on their phone—but can we take that low precision data and use our higher precision data to cut through the noise? We’re looking at using open data sets like CCTV to help us build a bigger and better picture of the weather even at the movement of trees for wind.”
For Rory and Brooke they hope to keep creating new and exciting things. “We’re working more with high-tech clothing brands,” says Brooke. Rory is continuing to improve chatbots and share Filament’s work with the world.
Looking to join exciting conversations like these? Workspace customers have the opportunity to attend our WBI dinners and a whole host of other exciting events. Or maybe you’re looking for a new space to take your company? Discover our Workspace Advantage. Find your perfect office space today.
Call us on 0203 930 4547 or use our contact form on the site.
Books and further reading
The Net Delusion & To Save Everything, Click Here by Evgeny Morozov
Ben recommended this author as an alternative look at how technology is shaping our world. “Morozov critiques what he calls "solutionism"—the idea that given the right code, algorithms and robots, technology can solve all of mankind's problems, effectively making life "frictionless" and trouble-free. Morozov argues that this drive to eradicate imperfection and make everything "efficient" shuts down other avenues of progress…”
Rodneybrooks.com – Robots, AI, and other stuff.
As well as a host of topics on the subject of robots and AI, Brooks expands on Ben’s point on how it will impact businesses and how we work.
The Exponential View
Curated by Azeem Azhar a strategist, product entrepreneur and analyst this regular podcast sees him talk with some of the leading experts in the field of AI and explore the future of technology.
“TechEmergence is where business leaders turn to understand how AI is impacting their business and industry – and what to do about it.”
Life 3.0: Being Human in the Age of Artificial Intelligence by Max Tegmark
Max Tegmark is an MIT professor and co-founder of the Future of Life Institute, whose work has helped mainstream research on how to keep AI beneficial. Tegmark takes us to the heart of thinking about AI and the human condition, asking some essential questions: How can we grow our prosperity through automation, without leaving people lacking income or purpose? How can we ensure that future AI systems do what we want without crashing, malfunctioning or getting hacked? Will AI help life flourish as never before, or will machines eventually outsmart us at all tasks, and even, perhaps, replace us altogether?”