Article contributed by freelance journalist Francis Hébert-Bernier
Freelancers: AI won’t change your world, except for…
Artificial intelligence (AI) is being invited into the newsrooms of various major media outlets such as the New York Times and the Wall Street Journal, according to recent reporting by Agence France-Presse.
Although it’s undeniable this new technology could change how we work, I’m not among those who believe it will revolutionize the journalism profession, and certainly not the jobs of freelance reporters.
What sources does AI use?
The main shortcoming of AI in terms of journalistic methods is that it’s impossible to go back and confirm the path taken to produce information.
Even specialists in the field who’ve studied specific cases are often unable to agree on what’s happening in the inner workings of these programs.
On top of that, if you ask an AI system directly where it got its information, it tends to invent sources from scratch, even coming up with fictitious experts and generating hyperlinks that go nowhere.
Being able to trace the source of any information back to its source is the very essence of the journalism profession.
Even though future generations of AI may be better when it comes to traceability — which I doubt, at least over the medium term — conventional search engines, trips to the library and especially good old conversations with the people behind the news are likely to remain the primary means of gathering and checking the facts for our articles for many years to come.
What role will AI play?
The developers of AI-based solutions are very aware of these limitations and that’s why the first types of tools that will be introduced into newsrooms are expected to be more like editing assistants.
However, a slightly more intuitive iteration of Grammarly or Antidote is rarely what we think of when we envision everything AI has to offer.
Instead, we think of automated text generators and, yes, they’ll likely be launched in the coming years. Hockey game and stock market recaps from the previous day are often mentioned as the first types of content that will be entirely automated. I would also include short articles based extensively on news agency releases and adapted to the media outlet’s style and format.
Nevertheless, it would be highly unlikely for that type of text to be published online in a fully automated way, especially considering the need to protect against lawsuits in the event of errors.
I would therefore bet that for many years into the future, AI-generated text will continue to be carefully edited by young journalists, the same ones who are assigned to writing those articles today. Unfortunately, the only thing that’s liable to change is that their day-to-day jobs will be made less interesting for the sake of achieving incremental efficiency gains.
How much to pay for AI?
When the technology prophets talk about AI’s possibilities and its ability to transform the job market, they usually omit an important detail: the price!
AI is an advanced technology and, in all likelihood, deployment will be very expensive for the media platforms that embark on that adventure.
Keep in mind that media organizations are facing a serious financial crunch after losing most of their revenues to those same tech giants who are now trying to sell them AI-based tools.
Even though we can expect major media companies here at home, such as La Presse or the Globe and Mail, to one day follow the lead of the New York Times, those companies rely on very few freelancers.
We work mainly in the world of magazines and other periodicals that do wonders with budgets that are far from being able to fund a technology project of that scope.
Moreover, given that many media organizations are offering their freelancers the same rates they paid 10 years ago, I think we’ll still be more profitable than AI for many years to come.
—
Francis Hébert-Bernier is a freelance journalist who sits on Copibec’s Board of Directors. After starting his professional career as an IT recruiter and history researcher, he turned to journalism as a way to better interpret the world around us.