If you can’t beat them, join them

Instead of resisting it, companies are brainstorming ways they can benefit from artificial intelligence.

Artificial intelligence is showing no signs of retreat. And while sci-fi movies suggest that it’s time for us to pack up and run, media experts are taking a slightly more optimistic stance. So, rather than renounce it, people are instead joining forces – with caution.

So says Gideon Lichfield, global editorial director at Wired, Harry McCracken, global technology editor at Fast Company, and Traci Mabrey, general manager of Factiva at Dow Jones, who each appear to agree on one central idea: generative AI (the kind that can generate text) should be a tool in our arsenal.

But the good news is that we’re not starting from scratch. Yes, generative AI is uncharted territory for most of us, but so was Google translate when it first came out.

Navigating the ethical landscape of generative AI and journalism

“Like a lot of places, I think, we use AI transcription tools for transcribing interviews. Sometimes you’re doing some reporting on a story from a country whose language [you don’t] understand, so you use Google Translate just to help you get a sense of it. AI has obviously been used in our backgrounds for a long time,” Gideon says.

The difference now is that this kind of tech is capable of replicating human-made content. And just like advertising, design and language, there needs to be guidelines when it comes to using it in reporting.

This can help publishers avoid the kind of journalistic mistakes some outlets have already made (for example, ‘this version has been substantially updated by a staff writer’). “It’s very important for us to set expectations and have some clarity – some transparency – about what we would or wouldn’t use AI for. Because I think establishing trust is really important,” said Gideon.

Today’s AI software can create sentences that look like they were written by humans and “that’s why they appear very convincing”, but “it’s not rooted in understanding of the world or factuality, it’s literally just copying sequences or using statistics to produce sequences of words that look like things that already are on the internet,” Gideon adds.

Why then, if generative AI is merely mimicking human writing, are we placing such importance on it? Because this skill means it can do the boring or mundane tasks that humans want to avoid. For example, when asked if Dow Jones would use AI to summarise a school board meeting, Traci responded with a resounding ‘yes’.

Generative AI can also be used to save time on tasks like coding. “Coders have really embraced [AI] and it is helping them be more efficient and effective, and focus more on the most important parts of what they do,” Harry explained.

But generative AI isn’t the answer to everything. Trying to brainstorm potential headlines? This new technology has got it. Looking to write a social media post? Do it yourself.

“You really need a human to evaluate what is actually compelling…we should be using this as a tool that augments human capabilities rather than replacing them,” Gideon asserted. “[AI] just produces fairly mediocre text that looks like everything else, and so it doesn’t stand out and it has no conception of actuality.”

Besides which, journalism cannot rely on a tool that has a tendency to hallucinate facts and figures. So, working alongside it is the plan, but knowing exactly how and when to do that is the play.

To get more digital marketing tips and expertise, download our full Marketing and media insights report here.