Dave Coplin of Microsoft
a. I think AI is going to have a massive impact on all businesses over the next few years. There are a couple of reasons for this, the first is that we are beginning to light up a vast amount of data about our businesses that will help us understand how our businesses actually work (as opposed to how we think they work). Crucially, this data has always been there, it’s just that historically it’s either been invisible or too difficult to capture. Now with a new wave of data tools and connected devices, we can create a much bigger picture of what’s going on and how it can be optimised.
b. The second reason is that the algorithms that make sense of all of this data are getting both more powerful and more accessible. These algorithms were once the province of big tech firms and were inaccessible to any organisation that didn’t have access to massive compute and brain power. These days, thanks to the cloud and to initiatives and products from companies like Microsoft and others, not only is the compute power and scalability available to anyone, crucially the algorithms are accessible almost at the click of a button. All you need to do is to subscribe to the AI service you want and then you can just point it at your own data. All the heavy lifting has been done – you just need to do the fine tuning.
a. This is a great question and in reality, we’ve got a bit of both. The reality is that the cloud enables incredible compute power access to incredible amounts of data. That’s pretty powerful. Companies like Microsoft are making the power of their algorithms available to their customers via a subscription model so in a sense, what we have today is a business oriented service that runs on your data.
b. But there are still some big steps we have yet to take to achieve the full potential of our vision for AI. The first surrounds the data we have access to and it’s that even now, we simply don’t have enough data yet. I know, I know, it’s hard to believe in an industry that has been talking about big data for almost half a decade, but the reality is that big data is only “big” when the sample size starts to approach “n=all”. Up until that point, we’re still going to end up with mistakes and approximations.
c. The other major hurdle we face is that the algorithms themselves are still rather “dumb”. Of course I don’t mean that on their own they’re not incredible feats of software engineering, but the reality is, they still have to be trained by humans. AI is only as good as the humans that train it and as such that means that AI inherits all of the flaws and unknowns that we humans possess.
a. I think the main challenges that most businesses face in unleashing the potential of AI and ML is that they simply don’t know where to start. Pablo Picasso had it right back in the sixties, he said that “computers are useless, all they can give is answers”. The real issue that most businesses struggle with is that they don’t yet know which the right questions to ask are. Instead we tend to get seduced by the hyperbole and chase after the “answer” of more technology before we really understand what we’re trying to achieve.
b. Smart business leaders will do a couple of things: number 1, is to figure out what the question that you are trying to answer is. Are you looking for greater revenue, more engagement with your customers, lower operational costs etc? Next, you need to figure out how you might feed the algorithms that could answer that question. Where could you get the data from? Do you already have it? Could you buy it from somewhere else? Could you generate it by slightly altering your business processes?
c. Once the business has understood the question and fed the algorithm with the right quantity and quality of data, the final challenge is for businesses to react appropriately to the outcome. The future should not be left to a battle of human vs machines, we humans still have talents that the computers are decades away from understanding (never mind replicating) and we humans must also appreciate that computers are far greater at somethings than we will ever be, the best possible future belongs to those businesses that can combine the best of human capability with the best of the capabilities of the machines.
a. Any industry that has access to a broad and deep range of data and has the courage to step away from their traditional (probably analogue) business model. Advertising is a good example and we’re seeing this play out in the programmatic buying of advertising space. Manufacturing and engineering are also good examples with some manufacturers now using their own data (and that provided by the industry) to create predictive models that save money, time and in some cases, lives.
a. AI is essentially at the heart of Microsoft’s vision. Our mission is to empower every person and every organisation on the planet to achieve more through their use of technology. The way we believe that people will be able to add most value to their lives (both in and out of work) will be through AI. AI will enable us to create new products that will fundamentally change the relationship between humans and computers. We’ve been doing it for almost a decade, with examples like the Xbox Kinect, Skype Language Translator and Cortana as significant examples of technologies that were either developed by or powered by AI. Microsoft is also democratising the power of AI for both individuals and organisations by making instances of it available within their products and services.
b. All of these examples put the power of AI in the hands of the people that have the most potential to affect change and deliver value in their organisations.
a. AI is at the heart of pretty much everything that Microsoft is working on, as such it’s actually really hard to think of any of the products that Microsoft currently make or are working on that will not have a deep connection to AI.
Dave Coplin spoke at the inaugural AI Summit in London on 5 May. The second, larger AI Summit takes place in San Francisco on 28-29 September. To find out more, and to join us at the Fort Mason Center in September, visit: theaisummit.com
Feature image credit: Flickr