The world is fast running out of fresh information, with fewer and fewer journalists and other qualified researchers to go out into the field to mine it.
Information (call it news, data, research) has a sell-by date. It is not infinitely recyclable.
Artificial intelligence (AI) programs like ChatGPT are fast depleting the existing deposits of information, and in the process killing off the jobs of the fieldworkers who have been responsible for harvesting fresh information for millennia.
Journalists have become an endangered species, with global job losses running at over 20 000 a year since 2008 – and this affects everyone on the planet.
In South Africa hundreds of jobs have been lost, with more at risk.
Titles closes
Media24 stopped printing five major newspapers in 2024, and many other print publications have closed, gone online (with skeleton staff) or are on the verge of bankruptcy.
This is in response to an 80% drop in average circulation of South African print publications over the past 10 years, according to the Audit Bureau of Circulations (ABC).
A study by Northwestern University estimates that one-third of United States newspapers closed between 2005 and 2024, averaging 2.5 a week.
The US Bureau of Labour Statistics predicts that employment opportunities for news analysts, reporters and journalists will drop by 3% between 2023 to 2033.
In the United Kingdom, over 300 local titles closed between 2009 and 2019.
News outlets are also closing or scaling down across Asia.
The same pattern can be found in Africa, with mass layoffs in Kenya and Nigeria, in addition to South Africa.
News deserts
Grassroots media are also closing around the world, leaving “news deserts” where locals have no reliable source of the news which affects them in their daily lives – such as road closures, planned new developments, and new weekend markets or entertainment.
According to the Association of Independent Publishers (AIP), 26 or 10% of its community member publications in Southern Africa have stopped serving their communities since 2016.
The number of publisher companies dropped from 164 to 147 over the same period, while print runs fell from 7.5 million to 2.5 million a month.
Why does this matter to anyone who is not an information harvester or has no ambition to be one?
What AI can’t do
AI cannot and will never be able to go into a rain forest and report on illegal logging or spend hours capturing the mating habits of the lesser striped swallow.
Neither will it be able to interview a managing director about what has made her company successful or tell the story of how a mother lost a child in a raging river.
AI cannot hold a politician responsible for delivering on the promises he or she made during the election season.
For me, the depletion of the existing lodes of information is felt most acutely in the business, trade and speciality media – the journalists and documentary producers who spend time investigating, analysing and then capturing their findings in print, radio, blog or television – each medium has its own place, strengths and weaknesses.
Exhausting the supply of data
AI cannot craft in-depth reports using fresh information gathered from the people involved and through personal observation.
Journalists often get their best stories from throw-away comments during an interview.
Which AI cannot do. You get the picture.
If you don’t believe me, ask ChatGPT to look out your window and tell you what bird is singing or if a cloud is covering the sun.
Algorithms such as ChatGPT and Perplexity are fast exhausting the supply of data. Every piece of information they scrape off the world wide web (i.e. purloin) is old.
Yes, AI may package it well (this author is not immune to using AI for initial research or to summarise documents).
But, there is no new knowledge or information.
Stolen images
Similarly, every picture that an AI “art generator” compiles consists of images stolen from the photographers who took them or artists who drew them.
Soon – if it has not already happened – the bots and algorithms will have stripped bare the existing deposits of information.
This is already evident when doing searches – the same information is being regurgitated.
Perplexity, which provides the sources of its information, will have the same references when a query is rephrased or expanded on.
Follow them and you will find that many are so outdated as to be fake news.
With ChatGPT one cannot delve deeper to identify whether it is hallucinating or just repeating outdated information or lies.
Social media is even worse as a source of information.
A newsroom will have checks and balances in place to verify facts and to protect the innocent.
Academia has its own rigours.
People posting on social media and their own blogs have no such filters.
It’s not sustainable
Meta has announced that it is phasing out its third-party fact-checking programme across its platforms, which include Facebook and Instagram.
Then there are the algorithms which censor the information that you receive.
This became obvious during the Covid-19 epidemic, where anti-vaxxers and vaxxers had very different feeds designed to support their points of view.
The same is true of lovers of cats and those who prefer dogs.
To circle back to why this all matters, it is because politicians, bureaucrats, business leaders, theologians, and you and I are having to make major decisions based on less and less new information and data.
It not only the loss of journalists in the field which is leading to the unsustainability of AI-driven information. Researchers and academics are also falling prey to the fallacy that AI is more efficient and cheaper than people.
Deposits of fresh data
There is still time (but only just) to ensure that there continues to be deposits of fresh data and knowledge for AI to mine and harvest.
We must find ways to pay the harvesters and gatherers of information, and it has to be attractive enough to entice a new generation to be part of the solution.
A study published by the UK National Council for the Training of Journalists in September 2024 found that journalists were, on average, older than the rest of the workforce.
Anecdotal evidence would point to the same being true in South Africa, despite the youthfulness of some newsrooms.
Pay for the information they pilfer
To start with, the ChatGPTs, Googles, Bings, etc of the world must be made to pay for the information and images they pilfer.
None, to my knowledge, have teams of information gatherers in the field, which means they are stealing other people’s stuff.
My local supermarket manager is totally unimpressed when I try to pay with “likes” or a following of thousands. They want hard currency.
Intellectual property regulations must be enforced.
Some governments and news organisations are making an effort, but it is frankly half-hearted.
It is also up to all of us to support journalists and researchers by paying for the information and entertainment they provide – be it through monthly subscriptions, purchase of physical publications or donations.
A new generation of publishers who break all the “rules” and put information before profits is probably needed to provide the value that people are willing to pay for.
Either that or learn to live with yesterday’s news and views – as nourishing as the stale bread information gatherers are being forced to live on.
Ed Richardson is a journalist and researcher who has experience in print, radio, television and online.
