What Watson did for the AI conversation

Reflecting on the impact of IBM’s supercomputer.

Sign Up

Get the New Statesman's Morning Call email.

The year 1997 was a defining moment for artificial intelligence. That’s when IBM’s supercomputer Deep Blue defeated the reigning world chess champion Garry Kasparov. And following that success, in the early 2000s, the company sought to accomplish what was previously thought to be impossible – building a computer system that could process and understand massive amounts of data in the realm of human language and knowledge.

The tech giant settled on developing a computer system named Watson – in honour of IBM’s founder, Thomas J. Watson Jr – that could compete on the TV show Jeopardy! When Watson again emerged triumphant in 2011, AI had officially entered the mainstream.

Project Management Institute (PMI) recently recognised Watson as one of the 50 most notable and significant projects of the past 50 years. The ranking highlights the ways in which project management transformed these groundbreaking ideas into reality.

Here, Jim Boland, IBM’s head of the Project Management Global Centre of Excellence, shares his insights into the development and evolution of Watson and AI...

What was the initial goal with Watson?

The first challenge was to develop a software application running on the right hardware that could digest a huge repertoire of information and sort through a wide range of questions to win a competition like Jeopardy!

The broader goal was to create a new generation of technology that could derive answers from unstructured data, in a way that is more effective than the standard search technology that existed at the time. IBM wanted to move computers from being back-office calculating machines and bring them to the front office, helping to improve decision-making.

How has Watson changed from its Jeopardy! days?

Watson has evolved into a suite of AI applications and open source APIs. Watson evolved from a specific application developed for a particular challenge to a very broad suite of offerings. By combing near-instantaneous data analysis, precise decision-making, and the ability to understand human speech, Watson can help solve a number of AI challenges

From a project perspective, was having the very simple — though not easy — and measurable goal of winning on Jeopardy! helpful in the successful development of Watson?

Like any large, first-of-its-kind project, you need to have a goal so that you can understand what success looks like. But at the same time, you need to, or should try to ensure that the foundations exist so that the technology can build into something greater and more impactful.

If you look at the version of Watson that appeared on Jeopardy!, it needed to have the networking connectivity from the box that was there in the studio back to the infrastructure. It needed to have the right architecture, the software running it, and the speed and performance.

And then there was the body of knowledge. To bring all those components together in a complex project was of course very difficult, and that’s why it’s critical to know what success looks like.

Do you think giving Watson a persona, if not necessarily a personality, helped the public become more comfortable with the idea of artificial intelligence?

Yes. For example, our Project Management Centre of Excellence recently developed a chatbot called Hugo for all of IBM to use. If anybody has a project management-related question across our 350,000-plus employee population, they can access Hugo and will receive the correct answer the vast majority of the time.

Hugo is now well known across IBM, and is a virtual member of our team. You can access Hugo 24/7 from anywhere in the world, via Instant Messenger, over e-mail, etc. All the ways you can contact a human, you can also now contact Hugo. And I think that’s becoming an industry norm now in relation to chatbots and the agents that interact with humans.

What are some of the challenges or situations Watson is being used for now?

We’re doing a lot of work with external customers in areas like knowledge gathering. Artificial intelligence is really transforming how knowledge is extracted and then made available in an intelligent way to workforces within a customer’s organisation.

A lot of organisations are looking at their most experienced employees potentially coming up to retirement over the next 10 to 15 years. The challenge for those organisations is taking all of that information and making it available to more junior employees or those who are just starting off their careers. That information has to be extracted, stored, analysed, and presented in ways to help those employees learn. Accordingly, it’s a tool to help organisations become learning organisations.

And then there’s customer care. How can we use AI to help customers get their jobs done quicker but also in a way that improves the overall customer experience? Customers want the right answer quickly. Sometimes that can be automated; other times it’s less than optimal to interact with an automated answering service or a chatbot rather than a human. You can integrate it with robotic process automation where it’s automating manual or repetitive tasks. There’s huge potential in this space.

In what other fields does Watson represent great potential?

There’s tremendous opportunity to make improvements in health using technologies like AI. Take life sciences, for example. Currently, only 3-5 per cent of patients are enrolled in clinical trials, and it takes an estimated ten years from initial discovery to bring a new drug or treatment to market.

 IBM offers solutions that help streamline clinical trial processes – not only helping patients to get on the right trial faster, but also shortening study build time and data capture. Our technology can help clinical research teams accelerate timelines, trim down costs and more efficiently launch and complete studies, bringing needed tools to patients sooner.

 From a research perspective, Watson’s doing a huge amount of work with some of the major clinics, academic institutions, and hospitals around the world, especially around Parkinson’s disease, cancer, and diabetes.

When the average person thinks of Watson, do they think about Watson being a Jeopardy! champion?

Watson was one of the founders of AI that is paving the way in which AI is applied to real-world challenges. And Watson, as important as it was, was one of a number of significant “firsts” that IBM has played a part in. We pride ourselves on our computer systems being instrumental in putting the first man on the moon. IBM machines even supported the first open-heart surgery. There are countless other historic events in which we have been involved from the start.

Jim Boland is IBM Project Management Global Centre of Excellence leader.