Artificial Intelligence (AI) is gathering pace as one of the biggest ICT trends of our time – and the topic of “robot journalism” provides important insight on how AI is being applied to industry.
It also raises a host of questions.
How is natural language generation (NLG) being used to explain big data in narrative form that broader audiences can easily understand? What insights are being found by machines producing 2,000 articles per second? How are some of the world’s biggest media agencies currently using vendors that provide these solutions to build their reach, engagement, and businesses?
Will “robots” (or automated content) soon render journalists obsolete? Or will it enhance their work, allowing them to please more readers than ever before and focus on what they do best: examining the how and the why?
The answer is most definitely the latter, says Joe Procopio, Chief Product Officer at Automated Insights, one of the leading companies in automated content with more than 200 customers, including Allstate, The Associated Press, and Yahoo!
“Artificial Intelligence is very good when you give it one thing to do,” says Mr Procopio. “It isn’t as good at multi-tasking.”
Mr Procopio – along with Prof. Joonhwan Lee of Seoul National University – recently presented their perspectives on “robot journalism” to ITU staff as part of ITU’s Emerging Trends initiative.
Both presentations laid out different use cases and how they are evolving to meet client needs. It sparked a rich discussion on a range of topics from the evolving technology, to the client use cases, to the ethical and legal issues that arise.
ITU News caught up with Mr Procopio outside the discussion for a brief Q&A via email. Here’s the exchange:
What are the main benefits Automated Insights provides clients?
Our clients derive several benefits from our technology. First, automation saves time and money when applied to existing efforts. Second, automation allows for the production of new, often highly customized forms of content that would not be possible otherwise. Third, telling a story around data makes that data more actionable and understandable. Clients may be looking for one of these benefits or some combination of the three.
How have you seen the market evolve in the past few years?
Our company focused first on automating stories around sports data. Then we began to apply our technology to other areas such as finance, business intelligence, marketing, and healthcare. In October 2015 we launched the first self-service NLG platform, allowing people to use their own data to automate their own content.
The market is starting to understand automated content, how and when it should be applied, and what its strengths are. A lot of misconceptions are falling away, and it seems to be a natural path that automated content will become a normal method of communication.
What’s next in the evolution of Automated Insights?
Our Wordsmith platform will continue to advance, making it faster and easier for users to automate content with their data. The amount of data that is in the world will only grow, as will the need to tell clear stories around that data. As we learn from our customers, we’ll continually make it easier for customers to connect their data to Wordsmith, explore the data and insights, and provide more varied, more complex, and more robust stories to their audience.
We’ve had interest in our technology from dozens of different industries, including e-commerce, finance, real estate, marketing, media, and business intelligence. We really can’t predict all of the ways people will use it. We look forward to being surprised.
Joe Procopio (@) is the Chief Product Officer at Automated Insights. Joe designed and co-developed the first version of the Wordsmith engine and continues to guide the company’s direction for furthering the science of automated content. He leads a team of data scientists, content architects, and technologists.