The Predictive Power Of AI
The Predictive Power Of AI
It is argued that the core benefit that AI technology will provide us is the ability to significantly lower the cost of making accurate predictions.
Join the DZone community and get the full member experience.Join For Free
Bias comes in a variety of forms, all of them potentially damaging to the efficacy of your ML algorithm. Read how Alegion's Chief Data Scientist discusses the source of most headlines about AI failures here.
When we think of the transformative nature of artificial intelligence, it's easy to get carried away into thinking it will touch every aspect of working life. No process will survive unscathed as AI disrupts the workplace from top to bottom.
Whilst it's a compelling narrative, it's one that doesn't have a great deal of basis in evidence, at least not to date. That's not to say it won't have a profound impact, however, it might do so in slightly more grounded ways.
It's a topic examined in Prediction Machines, the latest book by the University of Toronto's Ajay Agrawal, Joshua Gans, and Avi Goldfarb. They argue that the core benefit AI technology will provide to us is the ability to significantly lower the cost of making accurate predictions.
Inside the Winds of Destruction
Their insight into what AI is and is not primarily comes via the Creative Destruction Lab they created in 2012 to help accelerate the growth of science-based startups. During this time, the number of startups who have AI as a fundamental part of their offering has rocketed, but most have this predictive support at their core.
With this shift in our perception of AI, they hope to drive through not only a clarity in the debate around the best way to utilize the technology, but also in the direction managers can begin to deploy it in a serious manner.
That's not to say that this seemingly narrow focus on predictions diminishes the impact AI can have on our organizations. Indeed, that is far from the message Agrawal et al try and portray, as they see accurate predictions as fundamental to our success. By massively lowering the cost of such predictions, AI promises to not only improve the status quo but cause a rethink in how businesses operate.
For instance, one notable example they use throughout the book is the predictive shopping patent taken out by Amazon a few years ago. The theory goes that retailers will become so good at understanding what we want, they will be able to know what we crave before we do, and therefore ship products to us rather than waiting for us to order them.
Whilst the precise way that AI technologies might impact businesses are largely steered clear of, they do, nonetheless, provide enough pointers to provoke a new way of thinking about the technology among leaders. Given the number who still remain largely unclear on how the technology will impact their business, this is a good starting point.
As the authors are economists, they largely stay clear of the technical aspects of AI in favor of its economic aspects. As most leaders will be infinitely more familiar in this world, this is no bad thing.
"The new and poorly understood implications of advances in prediction technology can be combined with the old and well-understood logic of decision theory from economics to deliver a series of insights that help navigate your organization's approach to AI," they write.
It is undoubtedly a long way from many of the books written on the topic, but I would argue it is considerably more valuable than those that venture into the realms of science fiction. It's often said that whenever you want to change, you have to start from where you are, and today most executives are largely clueless about how and why AI matters to them.
As such, framing AI in terms of its predictive capabilities is not only a realistic portrayal of its capabilities today, but also one that business leaders can both understand and act upon. For that alone, the book is worth reading.
Published at DZone with permission of Adi Gaskell , DZone MVB. See the original article here.
Opinions expressed by DZone contributors are their own.