AI in the Workplace: The Peril and Promise of Advanced Computing

A Q&A with Clive Wilkinson Architects on the future AI in the workplace.

For the past three years, Metropolis’s director of design innovation, Susan S. Szenasy has led Think Tank, a series of conversations on human-centered design. On July 19, 2017, she visited the Los Angeles-based firm Clive Wilkinson Architects to discuss AI in the workplace. What follows is a transcript of the conversation, edited for clarity by S.T. White.

Susan S. Szenasy, director of design innovation, Metropolis Magazine (SSS): Let’s begin our conversation about AI with Frances’ encounter with Watson, IBM’s artificial intelligence system.

Frances Anderton, executive producer and host, ‘DnA Design & Architecture,’ KCRW (FA): I went to a computing conference, and when I arrived I filled out a screen with a few personal details, including my Twitter handle. I was led to a bar where a bar tender was following a recipe that Watson created according to predictions about my personality. The drink appeared, and it was gross. I felt mildly insulted.  Then, Watson predicted what my daughter would want for Christmas based on some details about her. It showed a black sweater with huge gold hearts, which I thought was awful. Later, I showed it to her and said, would you ever like this sweater? And she said, yeah it’s pretty cool.

The cocktail was a fun way for IBM to showcase the robot’s ability to crunch data, and to introduce Watson in a social setting to businesses.

SSS: Ashley, you’ve found interesting ways to utilize technology in urban planning. Would you describe your vision?

Ashley Z. Hand, AIA, LEED AP BD+C, co-founder, CityFi (AZH): If we can operationalize data at massive scales and be able to understand how systems work with and against each other, technology can provide useful feedback loops. We can predict where a pothole is likely to happen and respond with a redesign of the service delivery model to avoid the pothole until it’s repaired. We can be proactive instead of reactive.

It’s a fundamentally transformational rethinking of the way government works. There is a lot of resiliency potential and we would have an increased ability to respond with the right-size service. A field inspector equipped with a tablet and real-time data would be able to better understand the context of an incident.

SSS: What infrastructure is needed to become a “smart city”?

TBWA\Chiat\Day Los Angeles. © Benny Chan/Fotoworks.

AZH: Cities need a communications infrastructure, which is largely in place, but there is a significant digital divide that we must be sensitive about, because we might not be accessing an entire city. Crowdsourcing is not thorough unless the crowd represents everyone. Hardware and sensors have to be placed somewhere, on poles, underground, etc. Then, there have to be analytics, the software behind it. Finally, we need to organize human capital that manages the systems, turning data into real knowledge. We have a void in the public sector in providing knowledge and expertise around the systems themselves.

Geoff Wardle, executive director, Graduate Transportation Systems and Design, Art Center College of Design (GW): AI presents many opportunities to create a seamless, multi-modal transportation system. In Los Angeles, we have a number of subsystems, but with AI they have become more coherent and navigable.

FA: At what point does this accumulation of data turn into potholes getting fixed? Is AI accountable for that?

AZH: The power of AI, particularly around mobility, is the ability to think systemically. We can infer that groundwater is the source of a problem, for example, when we can assess a large set of patterns. There is power in the idea of “mobility as a service,” streamlining everything into a single user experience. What makes a city is not the buildings, but how people interact with the environment.

SSS: Could you talk about a specific project that incorporates crowdsourcing?

AZH: ATSAC, the traffic management center in LA, was originally created for the events of the 1984 Olympics, and has since covered the entire city.  Underground magnetic detectors respond when a car drives over them which uses analog technology to synchronize traffic lights. It happens in real-time and is estimated to save people 13% on the time of their commutes.

A new project is the Great Streets Initiative that will use sensors to measure environmental factors, like heat, air quality, and noise. That data will be overlaid with crowdsourced data from Google Waze. The updated data challenges assumptions algorithms have made using older data.

FA: Do these systems communicate with each other?

AZH: Google has made a decision, business-wise, that all they care about is optimizing the trip time. They don’t care if it’s a suicide straight or a hard left. Period. The City of Los Angeles provides all road closure information to Google Waze, but you still drive up to places the system doesn’t recognize. The information isn’t a live feed. It’s a PDF. There are still glitches.

FA: But it nails what the real AI issue is, which is proprietary information.

AZH: Yes, Google Waze will not share the full suite of information, which would help us adapt our infrastructure.

GW: The conflict between corporate interests and ours has long preceded artificial intelligence. With AI, there’s an opening for it to go further beyond our reach.  The more opportunities we have to question our relationship to AI, corporations, and our ability to trust them, the better. Data ownership is fundamental to this discussion as we continue to produce data for corporations every day.

AZH: The idea that technology is a panacea is a falsehood. As we think about smart cities, communities should come together to discuss their visions of the city. What do we want and how can technology help us get there? What do we, as members of society, forfeit in the social contracts that we have, like allowing smartphones to report information? Maybe I’d feel okay if I trusted that the information would benefit the public good. Now is an important moment to define the values of our communities and society.

FIDM. © Benny Chan_Fotoworks.

Clive Wilkinson, FAIA, RIBA, president, Clive Wilkinson Architects (CW): Technology seems to move so far ahead of us that we’re grappling to understand it. In architecture, we are noticing some significant shifts in clients’ needs. Twenty years ago, they were mostly large, multinational corporations. We had rigid formulas for coping with organizations that were based almost on industrial and military backgrounds. Now, organizations are scaling down. Sometimes, that leads to fewer employees and jobs, but technology also offers the incredible ability for small players to act like big companies. People with specialized skills no longer have to work for large organizations, because the equipment they need to offer their services can be encapsulated in a laptop. They can work in a coffee shop, and offer things that are radically smart. The physical landscape of offices is shifting to co-working and flex spaces. Designing co-working offices is very different than designing for Google.

FA: Are you already designing smart buildings?

CW: There’s huge curiosity but it’s still new. The user experience would become software-driven. There can be every imaginable IoT (Internet of Things) features. When a person walks through a building, rooms would light up, adjust air flow, etc.

SSS: What other hopes and concerns should we be aware of?

GW: AI can help improve quality of life, completing unpleasant tasks, thereby freeing us to do what we should do or like to do. However, we should be careful about feeling out of control and feeling that we are controlled by people and entities we dislike. It would be a problem if AI made decisions that we disagree with but are unable to contradict. Lastly, there is the issue of gainful employment.

AZH: I agree. We can use AI for worthwhile goals, like humanizing cities. There are massive jobs that individual people can’t do, but that if done correctly by AI, would provide a higher quality of life.

Audience Member: I’m a lighting designer in aviation. We are trying to solve the problem of, who’s going to make the decision in an emergency? Is AI different than automation? Do we want the aggregate of information to make a decision it hasn’t made before? If I am flying a plane, I don’t want it making decisions for me.  I would prefer there to be a human responsible for the decision.

GW: I’m under the impression that when I fly commercial the plane is on auto-pilot most of the time, and it works well. I think one in four landings has to be done manually. I’m not scared of AI making decisions in general, but rather my issue is which ones they make.

One Shelley Street. © Shannon McGrath.

AZH: Ethical decisions can’t automatically be entrusted to AI. My career path would not exist if it had been contingent on an algorithm. My resume was not always in the top ten, but a person saw something unique in me and gave me a chance. Patterns of human activity contain people-made biases, and AI carries that bias forward. I heard a woman from the previous administration of the FCC ask, how can we regulate something we don’t understand?

Audience Member: It seems like the train has left the station, and AI is on its way to becoming an integral part of our lives, creating simplicity in transportation, buildings, etc. I’m wondering why we shouldn’t fully embrace it? Does the future really contain so many dangers?

GW: If AI manages us in our daily habits, making a seamless existence possible, I’d worry that we might forget how to be critical thinkers and problem solvers. A lot of cars now park themselves, so how many people have lost the ability to parallel park? Will we become sort of goofy as doors open before us and lights turn on?

Audience Member: But isn’t it possible that we would develop new intelligences?

AZH: There continues to be a digital divide. Technology is not pervasive in every community, so when thinking about smart cities, I consider the parts of the city that are not connected. If a city becomes more reliant on data, which isn’t collected evenly, then parts of the city could be left behind even more.

Just over two thirds of Americans have access to a smartphone, but we don’t know how many have data plans. There are unbanked communities as well without access to credit cards. As companies move to digital currencies and terminate cash transactions, the societal divide grows wider.

Audience Member: It seems like the technology industry is expert at concentrating profit and advertising. How is that likely to manifest in designing cities?

GW: As designers, I think we should question the reflex to advance innovation. Designers have to ask, “We can do this, but should we?” Apart from AI, the whole system that we live in is rigged. We are capable of imagining solutions to problems at that scale–not just products and services– since the government and economic systems don’t work well for most of us anymore. Maybe we don’t have to accept the status quo of making a few people extremely rich through advertising and data while the rest of us struggle.

FA: In LA, one of our strongest technologies is storytelling through pop culture. It’s a good moment for our fears and fantasies to converge and create hybrid visions of a reordered system.

Szenasy (left) and panelists at the offices of Clive Wilkinson Architects. © Joshua Franklin.    

Frances Anderton, executive producer and host, ‘DnA Design & Architecture,’ KCRW
Ashley Z. Hand, AIA, LEED AP BD+C, co-founder, CityFi
Geoff Wardle, executive director, Graduate Transportation Systems and Design, Art Center College of Design
Clive Wilkinson, FAIA RIBA, president, Clive Wilkinson Architects

Susan S. Szenasy, director of design innovation, Metropolis magazine

The Metropolis Think Tank series is presented in partnership with Corian® Design, DXV/Grohe, Staples Business Advantage, Sunbrella Contract Fabrics, and Teknion.

Recent Programs