The forgotten dimension: how AI might bring context back to the newsroom

May 1, 2024
Truth lies in context, and context has been lost to fragmentation, superficiality, platforms and other things. Can artificial intelligence be the game changer for good?
notion image
Journalism and its history are often described as opinionated, and certainly, opinion is a significant part of the process. Journalists and publications carry their own opinions, and even when good reporting is done, the DNA of the product is inextricably linked to the beliefs and principles of the authors. However, there is one unsung, crucial part of the process that journalism does best. It's the construction of the scenario around each story to help the audience understand what is around the narrative and how it interacts with reality. If truth had to be defined by a word, it would not be 'fact,' but 'context.'
After the "Post-Truth" event, journalism has heavily relied on stating that facts matter and news organisations excel when they get the facts right. There are two perspectives here. The first is that if we are discussing the importance of facts, something bad happened a long time ago. Large portions of the audience don't listen to unfriendly media due to the crack made in our collective cognitive agreement. The second perspective is that facts matter less than context, or rather, facts don't matter right now because context has been missing for a long time. Polarisation requires context-poor news environments, and that's exactly where we are.
The decline in relevance of news media within the information system has ceded ground to other information actors such as niche sites and TV shows, social media influencers and celebrities, viral content organisations like Buzzfeed, and platform algorithms. This shift occurred primarily because these new players managed to accumulate large follower bases in the digital sphere, which is their main competitive advantage. These newcomers entered the game with a new rulebook in mind - to garner more impressions, clicks, likes, retweets, and followers. Accuracy has never been a primary objective.
This is why context is so often overlooked now. Context was essentially the tool of the trade when explaining daily facts, but it was truly a burden. Any journalist who has been part of a breaking news coverage knows how laborious it is to gather all that supplementary information that will never make it to the front page. Journalists did it anyway, because it was not up for discussion. In this new information era, doing something that demands hard work, won't get you social interactions, and might take you hours to produce is simply a poor decision.
This is the dumb decision that has been made. Context has lost its place because the economics and dynamics of the game have changed as the playing field has changed. Unfortunately, the game itself remains the same - report, inform, and contextualize. Here, it's possible to see the bottom line of this major crack: the task [inform] and the need [nurture a healthy society] remain the same, but the environment and incentives have wildly changed. Under the current terms, there's no way the puzzle will become whole again.
Stories without context are, at best, half-stories, or even outright lies if the author so desires. An ad created for the Brazilian newspaper Folha de São Paulo in 1987, which won a prize in Cannes (see it below, with translated captions available), discusses a great leader who is initially unseen. This leader took a country in tatters and transformed it into a powerful nation, all while having a love for art - and it turns out, it was Hitler. "You can tell many lies while only stating the truth."
Video preview
What's the role of AI in context? In technical terms, AI could be nothing short of miraculous. If properly implemented (i.e., ensuring there are frequent checks to guarantee that the datasets used are transparent, databases with supporting data like stats and studies are shared, copyrights are respected, and are free of any biases), AI could transform any article of, let's say, 2,000 characters, into a live infographic, with an amount of information large enough to make it undeniable. Each CMS would ship a small universe with each article produced with very little extra work.
The downside is equally significant. Unregulated AI providers and users could create fake or unreliable narratives so convincing that there may be no way for the audience to disprove them. For instance, in Brazil, during the pandemic, WhatsApp audios circulated claiming that the virus was a fabrication by big pharma for profit. These audios, allegedly made by high-ranking health officials and hospital directors, still circulate among Covid-skeptics. With AI-driven context, they could mislead large segments of the audience who might never stop believing them. As always, technologies are not inherently good or bad, but their use can be both.
Information systems are fragmented to near-subatomic levels. Although a certain level of fragmentation is beneficial for competitive markets, it also opens the door for disinformation weaponised by digital and social media platforms. This shows, once again, why digital information markets need regulation. If, instead of tech bottlenecks controlling traffic, we had separate federated platforms where information reliability was proven safe by design, adding a context layer would be relatively easy. All publication members would share common databases, and the information could be monitored. The power of this new technology has the potential to solve a serious problem. However, that's where it stops. It could do it; we can't predict what comes next.
Context is the body of information that puts brakes on this data-swallowing frenzy. Audiences accustomed to dealing with support information (such as "Explainers", profiles, statistics, and charts) inevitably have to slow down the process. The attention span seems to be declining (although the 8.5 seconds attention span is a myth based on a misleading article), which explains why visual material like charts and timelines are effective when a story is told. Creating interfaces that can provide context in a non-intrusive way is the first technical challenge for information actors. Developing CMS that allow the author to gather accessory information on the go is a second pending task.
Restoring context's primary role is a priority in order to isolate disinformation voices into a corner where their reach is smaller. It would be, in a sense, like reclaiming territory occupied by the enemy, because this enemy struggles under the light. It can't be fully accomplished now because if you don't have the trust of 70% of the audience, anything you say is useless, whether it's facts or context. Mainstream news media cannot drop their partisan flags because they fear losing the customers who pay their bills. So, it's clear that we cannot expect this system to recover, but we can expect it to resist and increase its reach among non-believers until a new paradigm firmly sets foot on the ground.
The combination of AI's massive potential power with emerging technologies that are yet to gain traction (like VR and augmented reality sets, or the new generation of wearables) is very promising. However, it can only be realized in a functioning society. It seems reasonable to suggest that our current role is to control the damage by imposing strict regulations and inspections on these new technologies to prevent foreseeable negative outcomes. The news media cannot fight this battle alone. It requires the support of policymakers, NGOs, philanthropists, and investors to start experimenting with new approaches. For the moment, bringing the battle to a standstill wouldn't be too bad, as victory will take a long time to achieve.

© Cassiano Gobbet 2023 - 2024