Robots are being rolled out to perform tasks once the preserve of humans across many industries — but they have not yet made a big impact on the newsroom. That may be about to change, as Raneem Hannoush discovered when she spoke to the AP’s Lisa Gibbs

Can artificial intelligence help to combat fake news?

There are tools that can be used to write story templates that then can be generated at a very high volume based on data. This kind of technology can be used by reputable news organizations to turn out basic stories on sports games or the weather or financial data, for example, but certainly, people who create fake news can also use the same story generating tool.

There’s other technology that I think that people are more worried about that can now take a series of images and put them together in ways that can create new fake images. Given how powerful the image is, there is a lot of concern that you will not be able to tell the real image from the fake one. We are working to build a tool that verifies posts on social media.

Have technological advances made fake news more serious?

The power of social media platforms today allows news to be shared very rapidly and you can see how something that is not true can spread so quickly, so I think that is one of the reasons why the problem seems bigger than it would have years ago. At the same time the technology does give new weapons to people who are interested in fraud to create news that mimics the ‘real thing.’ On the other hand, there has been a lot of discussion on how automation can help fact check and verify stories. Imagine reading a story on a website, and while you are reading you can see pop up information about the sources related to the story that prove that the data cited is accurate and comes from the source.

What does the “AP Verify” tool do?

It is a tool that we are building to help us vet social media posts faster. We are currently developing and testing it at our London office. It is being (supported) in part by a fund coming from the “Google News Initiative” in Europe. We are getting the support of a giant technology company to help us make the investment in technology that we need to experiment and develop a tool that is good enough. The media industry does not always have the resources it needs for innovation. Technology has always changed. In journalism, a hundred years ago it meant using a pen and paper and using a telephone to call reports in, and then things developed. Each era witnesses some disruptive technological change, but editorial standards must remain the same.

What’s in it for Google?

Google has decided that it wants to support high-quality journalism. It wants to support innovation in media. They have created funds to help pursue those types of projects. I think Google, and even Facebook and Twitter are recognizing that their platforms have done great things for being able to share news (high and low quality), but they are also aware that our industry has gone through disruption in terms of our financial sustainability.

Will you share it with other news organizations?
We currently have a lot of journalists devoted to verifying and finding social media posts from eyewitnesses to events, because it helps us tell the story and get the news out there. We do distribute to other media outlets as well. Building AP Verify is a 2018 project, so we are not there yet. Therefore, we are still not sure how successful it is going to be, and how long it will require to develop a useful tool. There are many questions that we do not have answers to yet, but I do think that there is a goal around helping other media adopt those tools and practices.

Will the tool be able to read Arabic?
I am guessing that we are focusing for now on building the tool for English. I would think that we would look to adapt it for other languages quickly after that. Metadata that is in an image like date and time are universal, so there are a lot of things that can be done irrespective of language. Arabic is a very important language to AP. It has two foreign language translation desks, one is Spanish and the other is Arabic.

Will AI eventually make us journalists redundant?
We (the AP) have eliminated zero jobs as a result of our work with automation and AI and the way we approach it is that this technology is aimed at making our journalists more efficient and about removing the basic low-level jobs that people do not necessarily want to be doing anyway. We want our journalists who have expertise and skills to be doing things that matter, like telling human interest stories, doing investigative work, using the skills that they have in the best way possible. We do not want journalists sitting there in front of the computer like robots, turning out very basic simple stories about the weather or financial data or last night’s sports score. These stories lose value quite quickly in the Internet world. We see that journalists should work on stories that have an impact and last.