We recently looked at how the Press Association (PA) is using artificial intelligence to help revive local news journalism in the UK. While this example and a smattering like it represent evidence that the industry is beginning experiment with AI, the technology remains a relatively new concept within the world of media (and to be fair in many other sectors).
There also remains, it has to be said, an apprehension about the implementation of such technology, particularly in an industry which has already in many cases seen a reduction in workforce numbers following its earlier transition from print to digital, and beyond to mobile and social.
However, as Thomas Ross, senior partner for IBM Global Business Services here explains, there remains a lot of conjecture around the future of AI. For now at least, it makes sense to focus on the practical capabilities AI provides, and the industry benefits it can bring.
“One thing we are looking at right now is helping businesses to distinguish between the AI that is ready there to use,” says Ross, “versus the sort of things that are very interesting to debate, but not ready for implementation yet.”
“Because AI is such a huge topic, and I feel sometimes from my conversations with clients that we sometimes don’t distinguish these two things. Like differentiating between the big AI and the general AI, and the fact that in the future we see AI becoming more intelligent than us. The truth is that AI is smart for very specific use-cases today, and it’s good to deploy those use-cases, instead of only debating what AI can do for us in the more distant future.”
While the deployment of AI may be relatively new in the world of publishing, for IBM it’s practically bread and butter. The 108 year old US tech company has remained one of the largest businesses in the world over the years, and lives and dies on innovation.
In 1996, the company’s chess playing computer ‘Deep Blue’ successfully defeated the then World Chess Champion, Gary Kasparov. It was a seminal moment in the advancement of AI, which brought artificial intelligence into mainstream consciousness, and goes some way towards showcasing the dedicated work that IBM has been carrying out in this area for years. As Ross here explains, by the time this example had hit consumer headlines, the company was already thinking about the technology’s next incarnation.
“One of the things I try to get across is that IBM as a tech company is very old, and we have always had a focus on both day-to-day business, as well as thinking about what is coming next down the line. Because without IBM doing that we would not have continued to succeed as a business. So if you remember back to the nineties for example, when we were playing chess against the world chess champions, the thinking then was ‘What’s next’, right? So if a computer can compute a lot faster than people, then what’s the next thing for computers?”
“IBM determined that computers understanding language, and content, and pictures, is the next big thing. So computers have always been good at doing large numbers of data, but computers understanding content is from IBM’s perspective the next big thing. And this is where AI comes in. You need AI concepts and technologies to help computers understand content. For example, they should be able to do your tax returns, which is something they still can’t.”
“To understand the biggest advantage that AI presents, let me draw on another analogy, with robotics. When we started to have robots, the first use was not the ballet school in the Bolshoi Theatre. It was in factories. And so this same way of thinking, in identifying tasks that are too tedious for us to do, can be thought of for AI. It’s very good at doing the sorts of tasks that keep us away from doing the more creative tasks.”
Despite the claims of IBM (and the headline of this article) that the technology is not about to replace the human brain, Ross is open about AI’s ability to drive down workforce numbers through efficiencies. Berlin-headquartered online retail giant Zalando, he tells us, has been able to significantly reduce marketing headcount through its implementation.
“We should be doing more of this in the media industry. There are a lot of other industries that have embraced deploying AI more, like insurance, banking, and throughout public companies. But in media we haven’t done that thoroughly. So if you look at somebody like the fashion retailer, Zalando for example, they laid off 250 people in marketing. Not because AI has become smarter than their marketing people, but because they just needed less manual work in marketing, and that work could be taken over by intelligent machines. So there’s big potential in doing these kind of things.”
How far morality should come into the successful implementation of AI is a subject far beyond the parameters of this article. However, it is clearly something that IBM is thinking about. The UK press recently reported that the company has been busy rolling out software to alert businesses if their AI systems risk becoming racist or sexist.
For now at least, we focussed on what the physical implementation of AI within media companies looks like.
“I mean you can very simply go to IBM’s website and use the APIs. All the technology is already out there and you can use it, there’s free usage tiers, and then you need to pay when you use it more. Or indeed, yes we can do joint projects. So I think the benefit is that we have both the technology, as well as one of the largest specialist workforces on AI, data modelling, data infrastructures, computer linking, etc. And of course the more interesting things tend to be where we co-create with the publishers.”
“We’ve worked for example with Bundesliga in Germany, which has the world’s largest archive of football content. They were unhappy that the people who license Bundesliga content do not effectively use that archive. And they do not effectively use it, because it is too difficult to search. So we used AI technologies to make it easier to find interesting, emotional scenes to be broadcast and put together with current content.”
Finally, we had to ask Ross how likely it was that machine intelligence was set to overtake human intelligence anytime soon…
“I think it’s going to be a long time before we are overtaken by machines! The human brain is a very versatile thing, which I think we are only just beginning to understand the achievements and complexities of. But the thing that we need to do today – as we move more decision making to machines – is to understand how those machines make decision for us, or on behalf of us. And that is also something which is a research focus of IBM: how do you create transparent AI? And how do your understand where it has a bias, or may be unfair? That is a very important discussion, which it’s very important to have openly, so that everybody understands how AI supports us and takes decisions on behalf of us.”
More like this