The document summarizes a BIMA breakfast briefing on robots. It lists the speakers which included directors from Mando, IBM, Goldsmiths University of London, and DigitasLBi. The speakers discussed topics like the explosion of robots, how robots can carry out complex actions automatically, how robots are given agency, and how they are better able to cope with new situations compared to humans. Things to consider with robots included habit, coordination, and accountability.
[Good morning]. Thanks for coming. We know it’s never easy to take time out of your day. You’re all busy and important people. But just how important, and will that remain the case over the coming few years?
We know that there is huge potential for AI to transform our lives, and to transform the relationship between businesses and consumers.
It’s impossible for one person to hope to cover all that’s going on this this world and certainly not in under 10 minutes, but I’ll attempt to do a whistle stop tour of some recent highlights and pick up a couple of aspects of AI that should matter to all our businesses.
Now to understand why AI is such a big deal now, you need to take a step back and look at the impact of the “iPhone decade” we’ve just finished living in. It’s been pretty transformative. We’ve seen a number of science-fiction predictions come to pass and that has set the bar VERY high for consumer expectations of what’s possible with digital. [Examples: I can make video calls, track where I am in the world, and know my heart rate and how well I sleep!). However, it’s also true that while we can create a very compelling individual experience (say on a device or at an event) what we haven’t been able to do it make that a experience coherent across different contexts and devices – because that demands a level of machine intelligence that just hasn’t been possible before now.
So what we’re seeing is all the big players, including these four, invest massively in AI research, either buying up companies that are progressing in this area, or building in-house capability.
Google = at least 12 companies I’m aware of (deepmind, dark blue) for AI, facial & emotional recognition, neural networks, natural language processing, …
Facebook = at least 4 but focused internal team (Moments app – describes image to visually impaired) - Mark Zuckerberg reference to Iron Man’s JARVIS
Apple = Siri platform, Emotient (attention/engagement/sentiment for facial recognition during digital experiences - neuromarketing)
Microsoft = Cortana platform, SwiftKey keyboard interface with predictive AI component
$B dollar+ acquisitions and investments happening in the last 6-12 months alone.
Why? Because they all recognise that this is the key battleground for customers over the coming decade.
If these businesses can move from very simple rules (If this then that) on how to engage with customer and in what way, and get towards interacting based on emotion cues from my face, context and what my diary schedule was looking like, nuances of language I’m using, wider social data, health metrics, economic indicators,…. then they have the potential to connect in a much deeper way with customers and own that space.
So what have we got to show for it so far (since 1956 and the term AI coined at Dartmouth conference)… Well, four quick examples…
In 2014 a computer program successfully passed one interpretation of the Turing test by convincing 10 out of 30 judges that it was a real 13 year old Ukranian boy for whom English is a second language. Even though this was a very flawed and limited approach, it’s clear we’re heading into territory where machines have much more sophisticated and comprehensive ability to engage with humans in a seamless and naturalistic way.
Google’s AlphaGo AI has beaten a few grandmasters at Go. Now you can’t win at Go by just crunching the numbers… you need to be able to build in much more visual intelligence that can then be applied to other areas (like planning different scenarios when running a fleet of driverless cars for example). Again, in itself not that huge a leap forward but another chipping away – another milestone.
Affectiva, Emotient and other systems – built using AI processing of faces and now available as a simple cloud service – standing on the shoulder of digital giants.
Would have been R&D budget of millions of dollars. Now available to allow people to map a monkey onto their face in an app.
And finally there have been some spectacular AI fails. One recent one was the beautiful but tragically doomed chatbot Tay.ai, a Twitter entity powered by a Microsoft AI that learned how to respond through Twitter interactions… because what better way to learn about humanity and the nuances of language than the Twitterverse. As you can imagine, the Internet rose to the challenge of educating this fragile young mind, and so in less than 24 hours we went from this [tweet 1] to this [tweet 2]. Thank you Internet for creating possibly the world’s first racist AI. Not at all worrying there.
OK, it’s not the robot apocalypse Stephen Hawking warned about but for a brand’s reputation it’s not far off!
That’s a quick skip around the current state of AI. We’re seeing some really interesting areas where human and machine interactions are beginning to blur. It’s by no means perfect, and we will get lots of reports in the media every time a Google car has an accident but the pace of learning means that we will see more and more of a transition of and blurring the lines by increasingly capable machines.
To finish, I’d like to just spend a minute looking at one particular area of concern with AI that impacts on brands. It’s not the ‘Stephen Hawkings predition of the total annihilation of the human race’, but for me the interesting area that’s emerging is this…
We’re seeing more and more instances where machine intelligence is being used to provide a compelling digital concierge service.
Siri – mostly PA system for your digital world
M – extending out to real world through things like ordering flowers, Uber, Amazon products…
Only a matter of time before the explosion of integration and specific services allows you to essentially have a digital concierge act on your behalf. As the experience gets better you’ll become more and more comfortable with this arrangement, and the ease of this experience outweighs the marginal loss in not being completely in control of that experience yourself.
We know for marketers, the primary attraction of AI is hyper-personalisation… that it can move us towards an audience of one. We know that people want a seamless and coherent experience, bringing in a whole bunch of personal context that we know people carry around with them online. That’s impossible to do without AI, because we can’t make sense of all that big data and we certainly can’t do it at scale across all our potential customers. But if we can, we know it’s possible to weave these experiences together in a much more compelling way.
But as different providers (Hey Google, Hey Siri, Hey Cortana, Hey Facebook) embed digital concierge this as more and more of an integral part of their ecosystems, they are creating a new challenge. If people can just ask their digital concierge to sort out things in conversation that’s great as an experience, but as a brand, think about the potential implications on our ability to market effectively. In the old world, if a customer saw something they liked online they would follow a link, see a product in context and maybe related things, be exposed to cross-sell and upsell, and be far more exposed to the brand message. In the new world, the customer sees your brand through the lens of a digital concierge. One implication might end up being that it’s far more important how easy it is for an AI to access your services, rather than the customer directly. In effect, we may have to become better at marketing to machines.
In this light it’s easy to see AI as some kind of existential threat. In the short-term it’s far more likely that AI will for the foreseeable future be a tool that enables us to augment our existing abilities and create far more compelling touch-points for customers. But we should always remember that the big players are in the middle of a major land-grab for this new territory to lock out the competition and create a new walled garden that brands need to pay them for the privilege of operating in.
Don’t get left behind – understand what’s happening in this world and start to use AI as a competitive weapon in your toolkit. Incredibly cost-effective (will hear more on that from the IBM team).
Thanks