Digital Innovation Digital Publishing Top Stories
4 mins read

AI in media: Myths and misconceptions

Is AI a magic technology that will make all our jobs obsolete? Or will it be the silver bullet that ‘saves’ publishing? There are plenty of myths, misconceptions and exaggerations around the use of AI technology in media. Peter Houston takes a look at some of the biggest in this extract from the new report, Practical AI for Local Media.

AI is not magic

The hype around the release of Generative AI applications has more in common with alchemy than actual publishing. And while the promises around these developing technologies are not quite as hollow as ancient offers to transform lead into gold, they are almost as fantastical.

There is no Philosophers’ Stone to secure immortality for local media organisations struggling with the day to day challenges of digital publishing. There is, however, a proven, practical set of tools to assist stretched newsrooms in their work to deliver content that local audiences will find useful and engaging.

Supernatural shortcuts might seem attractive but practical AI is a process. It sifts real-world data and shapes it into stories that people want to read, but the results are never perfect the first time. Trained journalists and editors need to do the work of training AI tools to deliver what their audiences need.

ChatGPT levels of inaccuracy are not an option for local media outlets producing content that people rely on everyday. Human oversight is the only way forward.

Cynthia DuBose, VP, Audience Growth & Content Monetization at McClatchy in the US, told us how newsroom staff are constantly training the bots they use to deliver automated content to make sure it is ‘fully valuable’ to their communities. “Our AI implementations are strong because of our editors,” she explained.

To implement AI effectively, publishers don’t need magicians on staff, just people that can teach the technology to do what they have always done.

AI is not a replacement for journalists

For exactly the reasons outlined by McClatchy’s DuBose, journalists are crucial to the effective implementation of practical AI. Practical AI is not a plug-and-play solution and newsroom input is fundamental to long-term success.

Yes, we’re hearing stories all the time about generative AI creating end-to-end content based on limited text prompts. The problem is that half of those stories are about how the technology got it wrong². ChatGPT levels of inaccuracy are not an option for local media outlets producing content that people rely on everyday. Human oversight is the only way forward.

Publishers are, however, using AI to create content quicker and at much finer levels of granularity than human journalists can. In the UK, PA Media’s RADAR newswire fuses AI tools with journalism skills to craft localised data reporting from national statistics.

The stories are often placed with local titles that Editor Joseph Hook suspects don’t have the staff to take on these types of trend and analysis stories.

Publishers can use RADAR’s stories straight off the wire, however, Hook also sees a significant minority assigning reporters to put wire content in a local context. “We see papers picking them up and finding a local voice, a local organisation, to do that last bit of the job.”

Smart publishers are using AI to double down on adding value, delivering unique data-rich information and analysis and comment from journalists freed up by automation. Publishers that see AI as a way to reduce headcount? They’re likely to miss the opportunity to distinguish themselves from the rest of the internet that Douglas McCabe talks about.

“AI is going to change how you think about your journalism. If the routine stuff becomes automatable… the onus is very much on what you can add. Can you add empathy, entertainment, insight, expertise, judgement, the human touch, creativity? All those things are going to be at a premium.”

Professor Charlie Beckett, Director, The Journalism AI Project, LSE

AI is not about technology

The mainstream press speaks about AI as if it is one single cutting-edge technology. In reality, it means many things to many people, as is pointed out in the AI Journalism Starter Pack³ produced by the Journalism AI team at the LSE’s journalism thinktank, Polis.

Charlie Beckett, head of the Journalism AI project, describes AI in journalism as a disparate collection of applications. The one thing they have in common is that they’re all trying to help journalists cope with an overabundance of information.

“It’s not about the technology,” he explained. “We’ve got people who are trying to create products. They’re trying to create tools, or very importantly, trying to see how that tech fits into a system that will use it in an efficient and effective way.”

Working within established tech stacks is crucial in making AI useful to newsrooms of all shapes and sizes. The most automated content is of little value if it can’t be delivered to the right people in the right place at the right time.

Cecilia Campbell, chief marketing officer at United Robots, explained that one of the most important parts of any AI project is getting content to where the publisher needs it to best serve the audience. “That could be apps or a tag in the CMS. You only get real value once it’s inside the publisher’s system.”

Campbell recounted a presentation she saw from a startup working in natural language processing. “Stop,” she remembered thinking, “Don’t talk about all the intricate functionality. Nobody cares. Talk about the problem. That’s where everybody should start their AI journey… Is there some problem that we can solve through this?”

To access the report for free, please fill in the form below.

Your details will be used to send you the Practical AI report, as well as information about future Media Voices reports and United Robots communications. Please note United Robots and Media Voices are joint data controllers for this report.

Republished with kind permission of Media Voices, a weekly look at all the news and views from across the media world.