Benjamin Mateev is CEO and co-founder of the digital news platform informed. In this guest article he deals with the influence of artificial intelligence on news consumption in times of social media, news aggregators and filter bubbles.
The world has a new news platform: Artifact, which has been available to the general public for a few weeks, is the latest work by Instagram founders Kevin Systrom and Mike Krieger. The platform offers users a news feed that displays algorithm-based articles from a variety of English-language media. The algorithm should be particularly smart and the app should become something like the TikTok of the news.
The business model is new and the AI is very likely better than that of the competition. But the basic idea behind Artifact reflects a development that has long determined media and news consumption: personalization based on personal interests, algorithm-based. This is how other news aggregators work, with minor differences.
The idea behind this development was once impressive: Users are shown exactly the topics and articles that interest them, others are filtered out. The more they use a particular news aggregator, the better the recommendations will be. At the same time, they are usually shown content from significantly more sources than they would consume without such aggregators and can thus benefit from the diversity of global reporting.
Unfortunately, the reality is often different: users lose trust in the media. You are overwhelmed by too much content, fake news and dubious sources. And they find themselves in “Filter Bubbles”, which in turn leads to one-sided views and polarization of society.
At Artifact, they say they pay very close attention to the quality of the news sources – while the articles themselves are selected and played out purely on the basis of AI, the media are curated based on certain quality criteria. A big difference to social networks like Facebook, where all publishers who do not violate legal restrictions are allowed to publish. In addition, what the publishers pay for is played out on Facebook, which of course does not speak for better quality. It remains to be seen how Artifact will align its business model.
Other news aggregators have also claimed to pay attention to the quality and trustworthiness of their sources – until content with Chinese state propaganda and right-wing blogs appeared in the feeds of many users. Many of the major platforms are now doing more than they used to to fight misinformation. This is a welcome development. However, it does not change the general trend, which is also being continued with Artifact: that content is being played out in a personalized and AI-based manner, with less and less human intervention.
What development do we see – and do we want it?
The fundamental question that we as a society and also we media professionals should ask ourselves is: Is this a development that we want to pursue further? Or should we rather look for alternative ways of designing the delivery of messages?
One thing is clear: AI has remarkable capabilities and development is impressive. It is happening so quickly that we cannot yet fully understand and predict what it will mean for business models in the media world. It is already certain that AI will never disappear. But that doesn’t mean that we should only rely on AI.
AI can quickly evaluate large amounts of data and derive a variety of insights from it, which is where it has a clear advantage over humans. Humans, on the other hand, have expertise and experience, are capable of creative thought processes, such as identifying certain patterns, and have their strengths where quantitative evaluations reach their limits. Accordingly, AI can do a lot, but real balance and “foresight” have not been part of it so far.
We come from a decade in which we celebrated the infinity of content on the Internet. We have increasingly consumed news in aggregators and social media. We are all familiar with the downsides of this: filter bubbles, misinformation, trust issues, news overwhelm, news fatigue… We are seeing that users are increasingly frustrated as a result and want a different kind of news app that gives them trust again instead of them to slay with content. With informed we built this kind of news app. Artifact, on the other hand, seems – according to the first impression – to continue on the other path of the last decade. This is about large amounts of content being thrown at you – paired with AI-based personalization.
The Instagram founders are back with news app Artifact
What does this mean for the news consumption of the future?
If you look at the possibilities and, above all, the limitations of AI when playing out messages, you can derive ideas for the future. For example the idea that we could partly consciously free ourselves from problematic aspects of the news consumption of the past years, which has caused so much mistrust of the media and the news over the last 15 years.
And also that more platforms and news services should be created where quality really comes first and where people also strictly monitor this quality. Now more than ever, we need to find the right balance between AI and human input.
And finally, it would also be desirable for people, and not (just) the algorithm, to take over the curating of content within such news services otherwise tend to be outside of their “filter bubbles”.
Those interested in the news would thus receive more balanced, less polarizing reporting and would receive a better overview of the media reporting from different sources, more context, more diverse perspectives. And would again be surprised more often with topics. At best, such a development could ensure that trust in the media and news landscape grows again – and at the same time the overwhelm caused by the large amounts of available content decreases.
Because when people can once again rely more on receiving balanced information from trustworthy sources and on the fact that not only technology, but experts have compiled this information for them, then this can ensure a new level of composure in dealing with the media. For a better informed and less divided society. So it’s time for a counter-movement to purely algorithm-based news consumption in order not to repeat the mistakes of recent years.