Transcriber: Elisabeth Buffard
Reviewer: Anna Sobota Sometimes truth is stranger than fiction. Let me give you an example
of one such case. I’m a product person,
and like a lot of product people, we ask a lot of simple questions without having many
of the answers ourselves. In my case, my questions
are about casinos. I’ve been to over 600 casinos,
not because I’m a professional gambler, although I can hold
my own at a blackjack table, but because I've installed, designed, and created products for the casino
gambling industry for 18 plus years. If you've walked inside a casino, you’ve interacted with code or hardware
that I’ve worked on. The word casino comes from the Italian,
meaning a small house. And ever since the beginning, these establishments have really
been known for the table game. A table game is defined
by a game of chance, usually played with some kind
of cards, chips or dice. And it’s no secret that,
since the beginning, the casino has tried
to watch this table game. In fact, in the old days, they used to have false ceiling with people sitting up there,
watching the game. Modern day casinos use cameras, and these cameras are utilized
for recording and monitoring all kinds of activities
on the casino gaming floor. In fact, more than likely,
the casino floor has more cameras per square foot than any other environment
that any of us have public access to. But surprisingly, until quite recently, we have been unable to embed
any kind of computer vision or AI to monitor what’s occurring
on the casino gambling table. Weird, no? You would think a place that has had
surveillance for hundreds of years is the one place that an AI computer
vision model would go first. But it turns out the real world
is a lot more complex than an AI model. So, what do we do? Well, there’s two things to consider here. Number one, the viewing angles. For as great as these cameras
are in the sky, are not necessarily capable
of embedding computer vision. And number two, even in a casino,
no one likes a camera in their face. So one of the ways we can go about this
is play the same kind of hand the Greeks dealt the Trojans
in the legendary Trojan Wars. And what I mean is embed it into something
that sits on the table already and use a Trojan Horse. What you see here is a device
that you may have walked by if you’ve been to a casino and never even thought of it twice
and not even looked at it before. But today and quite recently, these devices have embedded
technology in them that allow you to monitor the casino table in ways never seen before. Normally, you would be surprised
to know that a lot of these functions that are getting performed
are done by the human eye. There’s quite literally people watching
you wager at the table. Not anymore. And so, as we bring
these things to the market, we're able to augment data aggregation
using our technology, things that are usually done by people. So why does any of this matter? Well, just like prototizing anything, prototizing AI must try and create value
for all the stakeholders involved. And we do that in three distinct ways, in our case: with people, with risk and with reward. The table game staff is the most prominent staff
of people on the casino floor. And over many years,
we have inundated them with so many screens
and so much technology they spend more time
typing at those screens, than interacting with the guests. With cameras that are able
to do things better than people can, people can do things
better that people can. So we were able to elevate
this good friction. There's a famous professor
at MIT called Dr. Goslin, and she talks about good friction
and bad friction. And using AI, we can elevate
the good friction, the customer service, and downgrade the bad friction, the time spent typing things on screen. So, we’re able to augment the experience
for the player and the staff, under risk. Now, it's no secret that ever since
table games have existed, people have tried to alter
the odds in their favor. The famous MIT blackjack team
comes to mind, I'm sure many of you have heard of it. There might even be someone in here
who may have practiced some of these arts. Game protection is a very important
concept of the industry, and certainly cameras can view the table
a lot better than people. So again, we’re able to augment
the experience of the operator. And finally rewards. So the way the industry uses rewards, they use something called
a player rating system. And the player rating
is built on average bet. An average bet is determined
by your initial buy-in, your first starting bet
and the duration of play. And that determines important stuff. Basically what the casino
wants to comp you: your free breakfast,
your hotel rooms, etcetera. We can augment that with cameras so that the players are able to,
more efficiently, consistently, in a more fair manner, gain the rewards.,
instead of the human eye. So finally, next time
when you walk into a casino, you might think about the AI
that’s running in the background, the AI that’s aggregating data and essentially able to pamper
to your preferences in the future so it’s getting you your breakfast
and your show tickets more accurately. And generally speaking, as many of us are innovators
and designers, and we create products
and implement AI in the real world, we should think about minimising
the environmental factors and the impact of that AI
while maximizing the data aggregation and pattern matching it offers. And maybe, just maybe, along that journey, you might hire a product person
to ask the simple questions. Thank you. (Applause)
