Whatever your level of enthusiasm for Artificial Intelligence has been, now is the time to embrace learning to work with it. In this month of 'smartoctober,' we provide guidance - starting with an overview of the current state of the media sector.

Journalists continue to express concerns about Artificial Intelligence, according to a recent, comprehensive survey published by JournalismAI. The challenge of keeping up with the rapid evolution of AI was a key finding of the report, which consulted 105 news and media organisations from 46 different countries on the subject of AI and associated technologies. We'll be quoting the report throughout this blog.

Approximately 40% of respondents indicated that their approach to AI has not changed over the past few years, either because they are still at the beginning of their AI journey, or because AI integration remains limited in their newsrooms. More than 60% wonder whether the use of AI technologies maintains important values such as accuracy, fairness, and transparency.

_____________________

It’s smartoctober! Every year we spend a month deep diving into a subject or skill set that we deem to be absolutely essential to newsrooms around the world. This year it’s time to consider AI.

_____________________

Google search is also based on AI

During the presentation of the research results in Bonn, the report’s lead researcher Charlie Beckett did what has been desperately needed to be done and managed to convince even the most sceptical attendees that they are, in fact, already using AI, for instance when using Google during their research. "And you're going to use it a lot more. And every source or subject you're going to report about is going to use it." His summary: "You need to know about AI."

What we're truly waiting for are tangible, concrete use cases that stand out amidst the noise. This was a topic extensively discussed by news site publishers last week at the INMA Media Innovation Week in Antwerp. According to almost all speakers, AI is currently shaking up the media world.

Reuters, the news agency, recently echoed similar sentiments:

‘There is no doubt that AI can enhance the way news is reported and delivered, as well as improve workflow efficiencies for businesses across the board.’

When AI is introduced in an editorial setting, Reuters' Global Editor Jane Barrett poses three important questions:

  1. Can AI replace work that human beings are doing but may not need to be doing?
  2. Can it augment and add value to the work we’re doing?
  3. What role are experiments going to play in transforming our work?

Reuters is already attempting to use AI to detect fake news, including manipulations of photos, videos, and text, but there is a caveat as Barrett explains: “We are excited, and we want to experiment, but we want to experiment responsibly.”

AI is being used in news gathering, production and distribution

What are other media organisations doing? Around 85% have at least experimented with writing code, image generation, headline generation, search engine optimisation, and authoring summaries. When asked how media are using AI, this is the answer:

A few examples of tools that are being used:

  • Automated transcription, extracting text from images, and structuring of data after gathering
  • Applications that can sift through large amounts of data and detect patterns, such as data mining
  • Tag generation, notification services, chatbots and language models that assist in automating responses and extracting data

“Transcription services like Otter.ai are invaluable for reporters on deadline, and our tag tool streamlines production processes for editors”, so says one of the respondents. One other says: “CrowdTangle is one of the tools we use regularly. It searches various social media posts for ‘viral’ or talked about posts.”

And it’s therefore not much of a leap to see how editorial analytics tools like smartocto fit into this shift. Smartify was the first AI-driven feature introduced almost two years ago. Using historical data, AI can make predictions when specific actions need to be taken. Based on these predictions, the Smartify algorithm provides concrete tips on how to handle authors, stories, and sections at the very moment editors want or need to take action. They ask the feature what their best editorial steps should be, based on performance and timing. The response could be like:

  • Post article X on Facebook
  • Move article Y higher on the homepage
  • Create a follow-up to this article because visitors are still interested

What’s key - and what’s different - here is that interpretation of editorial data is translated into action. Artificial Intelligence takes on an intelligent human task.

The significant revolution of foundational models like Claude or ChatGPT is that AI is seamlessly applied to human day-to-day language. This opens up significant new possibilities for storytellers.

In previous years, analytics revolved mainly around human behaviour, measured in clicks, likes, and views. Now with generative AI, reporting systems know a lot more about what the content is actually about. Style, sentiment, readability, topic and format, a real match between audience and story qualities can be understood on a deeper level. This will no doubt lead to more specific personalisation, or help inform more advanced content strategies.

Erik - new

Erik van Heeswijk ceo @ smartocto

_____________________

What is AI?

Artificial intelligence is a collection of ideas, technologies, and techniques
that relate to a computer system’s capacity to perform tasks normally requiring
human intelligence.

What is generative AI?

It is a subfield within Machine Learning (ML), a subfield of AI in its own right,
that involves the generation of new data, such as text, images, or code, based on
a given set of input data.

Source: JournalismAI report September 2023

_____________________

Some media outlets indicate that they are already taking significant steps towards fully embracing or integrating AI into their workflows. One of the respondents of that survey is using GPT-4 to create summaries and translations of articles written by journalists for use on various platforms. "We are also experimenting with AI-generated images, headline alternatives, tagging articles, audio, and video production."

Perhaps it's also essential to mention what hasn't happened yet: automatic content production. Fully replacing human article production with AI is not only controversial but simply not the reality: AI’s not good enough to do that - yet. When the Dutch publisher Mediahuis announced the launch of an AI-driven sports site (resport.nl), many journalists raised their eyebrows. But what really happened? Articles from the press agency ANP were automatedly published online and given a headline generated by AI. And even that is overstating AI’s role: as Mediahuis was quick to point out, nothing goes to press without a human editor performing a final edit.

Meanwhile, the big bosses in the publishing houses are also delving into AI studies. Around one of the three respondents said they had an institutional AI strategy or were currently developing one. Charlie Beckett has an idea of what they are focusing on:

  1. Fact-checking
  2. Content personalisation and automation
  3. Text summarisation and generation
  4. Conducting preliminary interviews and gauging public sentiment

But we're not quite in the new era just yet. Media makers are facing technical and financial challenges, ethical challenges, cultural challenges, and managerial challenges, as evidenced by the results of the survey.

On that one question that raises the hackles of many journalists, Charlie Beckett had something to say in Bonn:

“Yes, it will take over your job, when your job is very repetitive basic work.”

So, he calls on every journalist who is hesitant to take the leap to educate themselves. Learn how to prompt well is his advice. And for those in charge: come up with ethical and practical guidelines. He even pulls out a step-by-step plan from his pocket for anyone thinking, "Where do I begin?"

If you continue to follow us during smartoctober, we will go through some of these steps. While it’s in our interest that you understand the developments at smartocto, hopefully, by spending some time with us thinking about AI and how it affects our industry, you’ll also feel more informed about the benefits and value that it has for journalism.