Skip to main content

What if rejecting anthropomorphizing AI isn't just an ethical imperative, but good business?

One of the most powerful arguments that Joanna J. Bryson makes in her (in my opinion, oft-misunderstood) Robots Should Be Slaves is:

 "(G)iven the inevitability of our ownership of robots, neglecting that they are essentially in our service would be unhealthy and inefficient." (Bryson, 2010) 

Azeem Azar's recent Exponential View interview with Kai-Fu Lee points to a distinct advantage in inverting Bryson's insight. 

What if recognizing that robots are in our service is healthy and efficient

Much of Bryson's insights underpin her belief, echoed in the OECD and other AI Ethics principles, that our interactions with AI and other robots must be transparent and that humans must be explicitly aware when we are interacting with artificial agents. This widely-held view is about avoiding harm, but as Lee points out, it may also have clear benefits.

You should absolutely listen to the Exponential View podcast episode where Azeem and Kai-Fu talk about many things including AI and robots, business, China as technology innovator instead of technology emulator, how American business approaches ignore emerging markets to their detriment, and vitally, how China's response to COVID-19 has accelerated the adoption of robots into human augmentation roles. (Actually, you should subscribe to the Exponential View. It's awesome.)

In the midst of COVID-19, China accelerated the deployment of robots to facilitate safe physical distancing. (Yes, I recognize that I'm skipping over all of the additional reasons why China accelerated the deployment of robots and used COVID-19 as a timely excuse.)

At 19:14, Kai-Fu Lee relates his experience of opening the door of his apartment to accept a delivery from a robot in his pyjamas. 

"It's quite convenient...," Lee says. "Now I can open my door in my pyjamas because there's not a human who will see my in my pyjamas, but merely a robot. And so I think our behaviours are changing." (Lee & Azar, 2020) 

Similar Chinese delivery robot, from the Denver Post

The robot doing the delivery is a below eye-level, wheeled, limbless, faceless, decidedly non-anthropomorphic design. I want to make this clear - it is specifically designed to appear non-human

Earlier in their conversation, Kai-Fu Lee tell Azeem that several factors went into the non-anthropomorphic design, ranging from costs, to design simplicity, manufacturing needs, to human comfort. With human comfort part of the design process, a benefit is found at the intersection of many systems and processes, which, in one way or another, puts the human at the centre. Centring the human is core to countering dehumanization that leads to slavery in the first place. (Bryson talks about this in great detail in her 2010 article.)

Bryson states:

"when I say “Robots should be slaves”, I by no means mean “Robots should be people you own.” What I mean to say is “Robots should be servants you own.” (Bryson, 2010) 

By designing COVID-19 era delivery robots as non-anthropomorphic servants, Chinese companies were able to drive their adoption in a time where robots' use can save lives by reducing the opportunity for pandemic spread, by reducing the chance that human users will feel judged by the robots when they use them. 

Non-anthropomorphic robots may reduce shame. Non-anthropomorphic robots free us humans from our projection of others' judgements of our own worth. These projections are both cultural and personal and have both behavioural and emotional impacts, often driven by our own complicated needs involving esteem and self-worth; shame. Relieving the pressure to appear "presentable" at the door to accept a delivery and thus avoid the judgement of another human during the midst of COVID-19 frees us, albeit momentarily, from servitude to our own self-judgement. 

The delivery robot doesn't care what you look like. That's healthy, efficient, liberating and good business. 








Comments