I spoke on a BIMA panel about AI and robotics chaired by Peter Trainor, with Jonathan Seal, Devika Thapar and Chris Brauer. Peter asked me to do my set-up presentation about robots; this is a summary of what I said. My talk (slides with my voice-over) is below. All the presenters’ slides from the event are here.
Update: I no longer work at Publicis Groupe.
What is a robot?
I like this definition:
Robot are things to which we give agency
I prefer it to the dictionary definition of “a machine capable of carrying out a complex series of actions automatically” because it draws attention away from the object and towards the point of it, and its consequences.
The technologies of artificial intelligence, including natural language processing and machine learning, are big contributors to an explosion of robots in the modern world.
Here are five non-technical things that are contributing to an explosion in robots:
They’re getting better at coping with the new
Getting things done in the real world is messy. The rows of robots we’re used to seeing in factories have traditionally depended on the whole environment being very controlled. Cutting-edge robots like Boston Dynamics’ Atlas cope amazingly with rough terrain and other changes in the environment.
One branch of artificial intelligence research is working on “artificial general intelligence” (AGI) which can cope with any kind of new situation. Google’s acquisition DeepMind, most recently famous for beating the best human players of the game Go, a year ago released software that learned to play a whole range of video games only from looking at the pixels on the screen. It’s like a baby opening its eyes for the first time and 24 hours later beating the best human player of Breakout.
We spend more time with digital channels
So much is obvious, but why does this matter? Because although we spend much of that time communicating with people, the fact that the communication is mediated through technology makes us more comfortable when some if it ends up being communication with machines.
Uber, for example, is really a robot transportation company. It’s temporarily using people to plug a gap that it hasn’t yet filled with machines. Its order for 100,000 semi-autonomous Mercedes S-class cars and its investments in research clear indications that when it can eliminate the drivers, it will.
We tolerate them more
People who have tried autonomous cars for any length of time report that, while initially alarming, the experience becomes mundane very quickly, and their desire to drive fades. And autonomous cars are not only robot presences in the world. Domino’s Pizza and Starship have produced robots that happily trundle alongside pedestrians to do local deliveries. Amazon and Matternet are developing autonomous airborne delivery drones. The EHANG 184 is a Chinese autonomous aircraft that transports you.
We find them more appealing
Robots for us to use are becoming cheaper, easier to set up and more reliable all the time. Robot vacuum cleaners (eg, iRobot’s Roomba) are practical and supremely helpful. There are window cleaning robots and robot lawn mowers in a similar vein. There is a even a robot that folds your laundry: even if it takes 10 minutes to fold one T-shirt, who cares if you’re not there waiting for it?
Infrastructure is improving
My favourite robot from CES 2016 was this one:
This washing machine from Whirlpool has a tank large enough for a whole 3-litre bottle of washing liquid. When it is running low, it orders some more from Amazon with its Dash Replenishment service. You can forget about checking: when a new bottle turns up, just pour it into the machine.
The much-talked about “Internet of Things” is an important enabler for connecting sensors and actuators in the real world. Many of these use proprietary networks to communicate to phones or other hubs in order to reach the Internet. Emerging recently are networks such as NB-IOT that allow low cost, low power devices to connect directly to the Internet.
Things to think about
What does this explosion of robots mean to you and your customers? Think about:
How fast habits can change. When any technology is cheap and effective it’s amazing how fast people can go from not thinking it’s necessary to forgetting they ever needed to worry about it. If you’re sceptical, remember life before your smartphone — if you can. Or, buy a robot vacuum cleaner.
The need for coordination between robots. The reason I like the definition of “giving agency” is the implication that you tell your robot what outcome you want, not all the steps needed achieve it. It means we see robots doing things on our behalf (see this extract from Microsoft’s recent Build conference showing how its assistant Cortana brokers conversations with other companies’ bots), and that means the robots will be coordinating with each other. Do you have a strategy for marketing to robots?
The challenges of accountability. If something is acting on your behalf, who is responsible for its actions? After all, you defined the goal, but it figured out what to do. There is much talk about this in the context of autonomous cars but the problem is much broader. Regulations and laws will inevitably struggle to keep up. Customers will need to understand who is accountable and trust them. What opportunities does this create for you? Which companies would you trust to act in your interest?