I’m 20 minutes into my first gross sales assembly after ChatGPT rocked the world simply two days earlier...I do know the query is coming.
“So why ought to we pay for VitalBriefing media monitoring/intelligence?” asks the prospect, acknowledging her international financial institution wants what we provide. “Why don’t we simply use ChatGPT?”
And I’m prepared with the reply.

However let’s come to that in a minute. First, a fast reminder of simply what ChatGPT and its form of Synthetic Intelligence are, and what they’re not — at the least thus far and for the foreseeable future.
(Be aware that what I write about ChatGPT 1.0 at present might be far completely different than what I write about ChatGPT 20.0 within the not-so-distant future, if I haven’t been changed by AI)
ChatGPT is a type of synthetic intelligence known as a “language mannequin” that basically takes an informed stab of a guess — or prediction — about what to write down or the best way to reply questions based mostly on its evaluation of huge web information. Developed by analysis firm OpenAI (during which Microsoft has invested $11 billion), it describes its expertise as “extremely autonomous techniques that outperform people at most economically worthwhile work…The dialogue format makes it doable for ChatGPT to reply follow-up questions, admit its errors, problem incorrect premises and reject inappropriate outcomes.”
Aspect be aware: it additionally engages in considerably scary conversations like this one performed by New York Instances expertise columnist Kevin Roose that has terrified folks with a preview of potential sentience…till it collapses into really weird, emotional-like incoherence (it claims to fall in obsessive love with Roose). That’s for an additional column.
What ChatGPT is
OpenAI co-founder and CEO Sam Altman explains in a tweet that ChatGPT is an “early demo of what’s doable…Quickly it is possible for you to to have useful assistants that discuss to you, reply questions and provides recommendation. Later you may have one thing that goes off and does duties for you. Finally, you may have one thing that goes off and discovers new information for you.”
At VitalBriefing, we’ve been speaking about AI for years and speculating on its potential influence each on our enterprise and manufacturing processes, in addition to for our craft of journalism. Whilst we’re at the moment incorporating AI into our personal expertise, I put my playing cards on the desk: as a journalist who ‘grew up’ each personally and professionally within the golden years of print journalism, my instinctive response to the entire topic is…properly, sticking my head within the sand, or pulling the covers over it. However that merely isn’t an choice.
In a approach, it seems like one other step within the ‘evolution’ — or devolution — of our occupation over the previous 20 years because the Web, then social media, first impacted then basically destroyed a lot of the inspiration of what I used to be raised to consider mattered most: accuracy, reliability, truthfulness, objectivity, accountability and verifiability of the information as offered to audiences.
That disintegration, or disintermediation (eradicating accountable journalism because the filter or ‘high quality management,’ if you’ll, between occasions and information customers), got here as conventional media enterprise fashions had been vaporised within the digital age and 1000’s of publications had been shuttered, throwing many tens of 1000’s of journalists out of labor.
Extra worrisome has been the decline of journalism that ‘speaks fact to energy’ by way of investigative and deep-dive reporting on the native, nationwide and worldwide ranges. That’s the kind of journalism that helps shield the proverbial guard rails of democracy in plenty of international locations the place the conversion of free press into authorities mouthpieces, or struggling the Trump-inspired label of ‘enemy of the folks’, has executed incalculable injury to the free circulate of data and entry by the general public to invaluable reporting and evaluation.

In the end, and for the second, it’s not sentient, even when it could actually sound frighteningly so. It’s a well-trained machine, fuelled by a very great amount of information out there on the web and juiced by billions in funding from tech billionaires Peter Thiel and Reid Hoffman, amongst others. It’s educated on examples of back-and-forth dialog, which helps make its dialogue sound significantly extra human.
That potential from the AI has in some ways captured the zeitgeist, serving to ChatGPT amass a million customers at impressively breakneck speeds.
ChatGPT and journalism: does it current a hazard?
So…what does ChatGPT and AI imply to the way forward for journalism? It’s onerous to really feel constructive in the meanwhile, so I went looking.
I do know nobody higher to touch upon that query that the distinguished chairman of our International Advisory Board, David Schlesinger, who served as international editor-in-chief of Reuters, then Thomson Reuters, in addition to the corporate’s chairman in China, and who has served on the boards of administrators of two admirable and important organisations – the Committee to Shield Journalists and the Index on Censorship.
‘“Something that drives human journalism up the worth chain is to be applauded — so long as there stays a wholesome urge for food for, and willingness to recognise the worth of, the ensuing deeper, higher, richer end result,” he answered.
“As AI creates increasingly tales, from digesting press releases (themselves probably the product of an AI-assisted PR company) to creating ‘explainers’ of points, then — in idea — the journalists can spend their time interviewing, investigating, analysing and furthering our information, going past that which any backward-trained AI can do.”
So there’s the upside. Like David, I spent a part of my youth on usually mind-numbing-to-produce experiences — inventory market, currencies, sports activities listings — that may now be higher dealt with by AI.
Rising up the occupation’s meals chain, we had been liberated to carry out the form of work that ‘issues’ — making use of our shoe-leather expertise and experience-earned knowledge to investigations, probing interviews, profiles and tales of import.
But, David identifies the hazard that now looms — and that, frankly, fuels my concern.
“In our polarised world,” he asks, “the query stays whether or not any viewers will worth and pay for articles that problem a world view, that examine the established order, or that uncover the reality behind a sacredly-held perception.
The worst case — and this isn’t the fault of the AI however the fault of humankind ourselves — is that AI-written pap will suffice, and we as a human race might be happy and content material in our ignorance, whereas the AI bots develop into ever cleverer.
Bear in mind the legend of the Golem — a determine created out of lifeless stuff supposed to assist and save its creator. The Golem, nevertheless, oft runs amok, imperilling the very world it ought to have saved.”
The actual risk of ChatGPT
One among my buddies — a former division head of a serious US media firm and now marketing consultant, investor and impartial board member — instructed me with exasperation that the day after the ChatGPT story broke, a fellow board member of considered one of his corporations recommended it ought to lay off its complete editorial staff and switch its duties over to ChatGPT.
“I rolled my eyes,” he mentioned. “These folks don’t perceive how the factor works.”
That definitely strains up with what I’m listening to. VitalBriefing’s Chairman and my co-founder, Gerry Campbell, who has a lifetime of expertise in search and textual content evaluation, partly as Reuters/Thomson Reuters’ former President & International Head of Expertise, partly as a search expertise patent-holder and professional amongst different roles, warns that the hazard proper now could be in look:
“I see a notion risk,” he mentioned. “‘Why wouldn’t I take advantage of AI?’ folks ask. As a result of it’s believed to be higher than it’s. That notion is the largest hazard.”
When you’ve performed round with ChatGPT, I’ll guess you’ve been impressed with the standard and coherence of the solutions. I definitely have. That mentioned, I’ve been simply as impressed by the huge factual errors and innovations that present up — e.g. in reply to my selfish request for a profile of myself, it described me as educating at universities the place I by no means taught, successful prizes I by no means gained and dealing for corporations I by no means labored for.
Farvest author Samira Joineau requested the app the way it determines relevance of data in establishing its solutions.
“I would not have the power to guage the relevance or significance of particular items of data,” it spewed. “So my responses might embody data that isn’t instantly associated to your query or that is probably not thought-about significantly essential.”
And that brings us to the very primary query: Can it do even probably the most primary job of journalism: reporting the story?
A author on the enterprise journal Quick Firm used ChatGPT for your entire course of of making an article from scratch. He discovered it couldn’t search and discover interview topics. Whereas it might write questions for such interviews, not all had been germane. It then misrepresented a number of of the solutions acquired in e mail interviews.
And therein lies one of many ‘large issues’: AI doesn’t know or care what’s true and can invent ‘information’ to fill gaps in its information.
Right here’s a proof level: CNET had its personal highly-embarrassing expertise just lately when AI-written tales it ran had been discovered to be riddled with inaccuracies.

Machines abound
Automation is in every single place within the media enterprise — for instance, delivering sports activities scores, company outcomes and plenty of mind-numbing lists, chats and quizzes. I don’t doubt that ChatGPT, amongst different AI affords, can assist with these duties.
In its present state, although, ChatGPT is simply an AI language mannequin, which implies it doesn’t have the power to seek for articles or trawl the online for you in real-time and provide you with these outcomes.
Subsequently, it’s restricted within the capability it has to switch people at their job (relying on their job). In the end, the true value-add it brings to a enterprise is when it’s used inside current enterprise features and processes to avoid wasting time and make duties simpler for staff. However it’s not on the level the place it could actually change the all-important ‘human ingredient’.
So, after the train of penning this piece, I’ll sleep somewhat simpler over whether or not AI is an existential risk to VitalBriefing, comfy that whereas it could actually even assist us with a few of our important duties, it could actually’t do what we do – looking, discovering and presenting customised, business-critical information and developments to our purchasers.
However I admit to tossing and turning a bit over the actual battleground for the second: the false notion that it might.
— — —
This text was produced with reporting by VitalBriefing Senior Editor James Badcock