AI + Journalism + Future

Have truth and reality and facts lost?

Ok, monthly mind dump time. No AI in this one, I promise, just me typing away, although that is what we must consider, given the evolution of the global whole. Except the images, which are all AI generated on the theme of journalism and AI and the future. As well as the tech or the jobs on an (inter)national economy level over the next few years, it has implications for everything related to content creation, media, entertainment, analysis, knowledge, fact checking, reality, truth and the very fabric of trust that currently governs the (global) public sphere. As we move into a world in which the generative AI tools are so good that your own kids or wife are not going to be able to tell that it’s not you in the fake images and videos and audio, where does that take us in politics, the stock markets or even the courts? The new tools are wonderful for anything creative but a disaster, or at least a new historical challenge, for anything related to reality and facts.

That, of course, means journalism and current affairs, as one form of content that might or might not grab someone’s attention, even within a news niche (“global”, “tech” or “Spain”), for a few minutes as they flick mindlessly through their 10 on-demand TV platform catalogue offerings with thousands of available series to binge on, check Twitter on their tablet and are interrupted at the same time by notifications from WhatsApp groups or news or weather services they have on their phone for some reason. That is the attention psychology state of affairs in January 2024 before generative AI hits the mainstream and floods that global public sphere with orders of magnitude more. Adobe reported in October that even during their first few months of testing with Firefly and AI things in Photoshop or Illustrator, their users had rushed to generate 3 billion new images. OpenAI is up to 100 million users a week. Microsoft only put GPT-4 on everyone’s smartphones over Christmas. Zuckerberg only announced he was to spend $15 billion on Nvidia GPUs last week, in a quest for some super AGI thing to beat Microsoft and Google.

The most gobsmacking example since I wrote to you at the end of December, just a month ago (time flies!) is surely Milei’s speech at Davos, in Spanish but translated and dubbed into English by an AI tool hours later, with his lips synced as if he himself were really speaking English. And it is believable. A couple of week ago, the IMF admitted that up to 60% of jobs in advanced economies could be affected by this. Microsoft’s new “Future of Work” report, published at the end of December, dives into more details: inforamtion work, critical thinking, creative work, managment, collaboration, communication, organisation, all going to be affected: “People took 37% less time on common writing tasks […] BCG consultants produced >40% higher quality on one simulated consulting project […] Users were also 2x faster at solving simulated decision-making problems when using LLM-based search over traditional search”. It’s not hard to imagine where that goes in terms of global jobs and business: workers everywhere will be pressured to adopt AI tools to double or triple or quadruple their own productivity or be sacked along with legions of their less enthusiastic colleagues.

A new attention environment at one end and new technology to learn with new pressures in global jobs, work, business and value creation at the other. Another paradigm shift for whole areas of human life in the 21st Century.

How might journalism or analysis or commentary evolve? Honestly, I don’t know. We are headed to a strange place. We might posit that a personal human voice, an emotional face speaking on a video blog, could be a better format, for example, to cut through in that attention environment versus the automatic, algo driven platforms or the coming generative AI tsunami, but the Milei example from Davos already shows that the tech will be good enough for any journalist or commentator, even supposing no bad intentions, to train his own voice and video AI and then he could just sit in a bar typing out what he wants it to say, and most of the typing and probably even the thoughts in the article or video blog will be done by GPT5. It will even be possible to switch out the background so that it looks like your favourite correspondent is somewhere he has never been. That brings us straight to the trust and truth problem online, not only for text but with moving images and sounds that we are seeing and hearing and that feel absolutely real and authentic.

Want a summary or a synthesis of a new document? Upload it to a GPT and chat to it. Want to understand what the Chinese or Russian politician just said? The AI translator-dubber can already solve that for you.

And that’s just imagining one well-intentioned good actor trying to use AI tools to do his job for his readers better. What about the bad actors, the media companies, corporations, intelligence services, criminals and political parties, who will wish to harvest millions or billions of collective attention minutes a day? Soon, it appears that they will plausibly be able to spin up whole regiments of such AI-generated, absolutely realistic avatars to argue in favour of whatever it is their two or three AI supervisors want them to, in line with government or party or company ideology and culture. Readers and viewers are going to get battered and, just like now but even more so, by the time a brave journalist somewhere, if that role continues to exist, has spent a few hours or days attempting to fact check the latest AI generated political or corporate hit job, the video itself will have been seen by millions of people in the intended target audience and done its job. Even if they do it right, who will care about the journalists’ reporting anyway, if they are from hated, elitist mainstream media? It will just be another version in a an ocean of attentional sewage. Perhaps newsletters like this one will provide something of an answer, some guidance, if the writers promise not to use AI but, as you know if you’ve tried the latest GPT, it already comes up with thoughts and ideas and views that you hadn’t considered, and you will have no way of knowing if I am keeping to my promise of humanness.

I have been experimenting a little with OpenAI’s custom GPT builder and it’s easy enough to get it to spit out left- and right-wing 250-word oppoising polarised opinion articles on today’s news. It even does snark and insults. Takes about 60 seconds and the output is…pretty good. I’m sure that with some more tweaking of the prompts, both for ideology and for writing style, it will be possible to obtain quick columns that readers will be unable to identify as AI. It’s already that good. You can also easily turn them into Twitter threads or comments, and spin up some AI X accounts with AI profile images leading to plausible sounding AI generated websites where you can download or buy or subscribe to AI generated information or creative products, which will further drive your ideological affiliation towards whichever tribal position you prefer. All already possible. AI fans will hope the new tech is only used for good, but nothing in human history suggests bad actors will not use it for their own evil ends too.

And we haven’t even touched on news gathering, on trying to report on and describe and send dispatches from some real world, some meatworld, event somewhere, whether it’s a conflict or a court case or a political protest. This process never takes less than at least a few hours and, depending on the story, could easily be several days, weeks or months. Even the best funded global reporting efforts on the biggest global stories still just produce reports that end up as blips on people’s flooded attentional radars. We have seen with the ongoing Israel-Gaza conflict that the information war or the propaganda battle, our very understanding of whatever might really be happening on the ground, all the way up to the recent ICJ case hearing brought by South Africa for genocide, has taken place thanks to the new technologies, social and mobile, despite Israel still blocking the 2,000 foreign correspondents from crossing into Gaza. Even at that level of major global story, journalism as a concept in this new world has been found wanting

I will keep sending you weekly thoughts here on The Spain Report about how the country is changing, I will aim for Sundays, but honestly I don’t know where we all go from here in terms of more content and not overwhelming everyone. The content strategy from Christamas is already not going to work in this new environment, I fear. Time to experiment and adapt. You are still interested in Spain, and I am still here. Somehow, we will meet in the middle.

• Understand a changing Spain better

• Guarantee the independence of this analysis

• Support original, human thought