AI-Generated Storytelling: Opportunities and Challenges (with Nigeria in mind)

AI-Generated Storytelling: Opportunities and Challenges (with Nigeria in mind)

By Azu Ishiekwene, a paper delivered at the All Nigeria Editors Conference (ANEC) on September 8, 2024, at Yenagoa, Bayelsa State, themed: Economic Growth and Development Strategies in a Resource-Rich Country

One year ago, on November 15, 2023, the Guild of Editors invited me to share my thoughts on “Nigerian Media, Sustainability and Existential Threats by Big Tech” with colleagues. Being asked again this year is a privilege, but I won’t be surprised if this is my last invitation.

Perhaps I won’t need to come as a presenter next time. A learning machine, let’s call the avatar Anaba, might have been developed to make the presentation. This may sound incredible, but increasingly, it seems that what AI cannot do does not exist.

In its most basic definition, generative artificial intelligence (AI) refers to computer systems capable of performing complex tasks that, historically, only human beings could perform. Journalists, for example, used to think of themselves as the masters of storytelling in a hurry – God’s gift to the world as gatekeepers. We’re humbler now.

Let me say very quickly that the widespread use of AI is not only causing anxiety among journalists. Other professionals, especially the Luddites among them, are very concerned, too. Recently, I wanted to redecorate my apartment. I asked a furniture company in Abuja to recommend an interior decorator. The two recommended insisted on a pre-inspection deposit of 100k, which I wasn’t prepared to pay.

What did I do? I went to Chatgpt (paid version) and imputed a description of my apartment, asking for a photo design for each space!

Chatgpt delivered it to me in minutes, complete with a floor plan and carpark and asked if I needed optional designs! If you tried this on Midjourney, not to mention augmented reality (AR), you would get incredible decor options in minutes!

Let me return to storytelling, which is why I’m here. Man has always had concerns whenever there has been a technological change. As Yuval Noah Harari said in his book 21 Lessons for the 21^st Century, “Fears that automation will create massive unemployment go back to the nineteenth century.”

According to Digital News Project 2024, “Journalism, Media, and Technology Trends and Predictions 2024”, by Nic Newman, here’s a summary of how media leaders (300 digital leaders from more than 50 countries/territories) view this year:

• Just half are confident about the prospects for journalism this year, with concerns about rising costs, declining advert revenue, and slow subscription dominating their anxiety.
• Almost two-thirds are concerned about a sharp decline in referral traffic from social media, especially Facebook and X (formerly Twitter).
• More publishers say they’ll be putting more effort into WhatsApp and Instagram after Meta’s decision to open up broadcast channels for publishers.
• Publishers plan to create more videos, podcasts and newsletters.
• More inspirational human interest stories, especially with the growing fatigue in negative news.
• Using AI for backend news automation is considered the most critical use of the technology by publisher respondents.
• Experimental interfaces to the internet include AR and VR glasses, lapel pins, and other wearable devices.

The salience of publishers’ concern about the opportunities and challenges presented by generative AI is evident in the forecast. I imagine a similar mood may have informed why editors made interaction with technology a subject of discussion at two straight conferences.

What’s AI up to?

Let us look briefly at two recent examples of the use of AI in storytelling, one in North America and the other in Europe, that have resonated in many parts of the world.

In Mexico, Grupo Formula, the country’s leading broadcasting group with 2.3m YouTube subscribers, created three avatars—NAT, SOFI, and MAX—three robotic journalists who generate content in entertainment, sports, and politics for the company’s social media handles.

The group’s director of technology and AI infrastructure told the Reuters Institute for the Study of Journalism, “The news stories that NAT, SOFI and others present are small stories and very focused towards young people who don’t connect well with the old-style newscast. We are looking to connect with these young people using technology.” Grupo Formula’s subsidiary, TV OAI, is the first news channel in Latin America powered 100 percent by AI.

More recently, a Polish radio station, Radio Krakow, announced the relaunch of OFF Radio, the first experiment in Poland where AI-driven characters take on the role of traditional journalists.

In response to concerns about the increasing role of automation in the physical and cognitive spheres, Harari said in his book, “It would be madness to block automation…in order to protect human jobs. After all, what we ultimately ought to protect is humans – not jobs.”

What opportunities exist?

The telephone bridged distances and made communication easier. The typewriter, the camera, the pager, the Walkman and the tape recorder helped significantly transform journalism and creative writing, whatever the concerns or disruptions when they were initially introduced. What opportunities does AI present, and how might journalists use them for storytelling?

• Streamlined content production: AI tools can significantly streamline content production, allowing journalists to focus on the creative aspects rather than time-consuming tasks and drudgery. For example, algorithms can assist in generating ideas or first drafts.
• Personalisation of content: Algorithms can analyse user preferences and behaviours to create bespoke narratives with individual audiences.
• Enhanced engagement: AI can create immersive and interactive storytelling experiences. For example, games and multimedia stories that adapt based on user decisions can attract individuals who may not typically engage with traditional forms of storytelling.
• Creation of new opportunities: The fusion of AI and human effort can create new possibilities that were previously challenging to imagine. One good example is the collaborative effort involving 400 journalists from 80 countries sifting through 11 million documents and 2.6 terabytes of data during the Panama Papers investigation.
• Generative AI can repurpose chaos: In an article entitled “AI news that’s fit to print,” Zach Seward wrote, “Faced with the chaotic, messy reality of everyday life, LLMs (Large Language Models), are useful tools for summarising text, fetching information, understanding data, and creating structure…but always with human oversight.” This article also shares some of AI’s best and worst use cases.

Challenges of AI use

• Ethical concerns in AI storytelling: Potential for generative models to create misleading information, such as fake stories or imageries that blur the line between reality and fabrication. The Cambridge Analytica case and COVID-19 are good examples.
• Quality and coherence: Often, AI can generate good content, but the creative flow and coherence could be a problem.
• Creativity and originality: Over-reliance can lead to losing personal touch and creativity in storytelling, although improvements in biometric algorithms are increasingly making this less of a challenge.
• Fairness and bias: Ensuring fairness in AI-generated narratives is a vital but challenging aspect of development. Continuous evaluation is necessary for identifying how errors are distributed across different population segments, as biases may emerge over time.
• Adaptability to genre and audience: Adapting AI-generated content to fit audience genres and audiences presents a challenge. Current models may need to grasp the subtleties of genre convention and audience expectations.

Limits and success stories:

AI is a work in progress, and we could use lessons from some good and ugly examples of its application even in countries where automated storytelling appears to be well-established already. First, the ugly experiences:

• CNET and the tech error soup: Last January, CNET, a tech website, was publishing stories – financial advisory – on short-term saving instruments, how to manage and close your bank accounts, etc. Even though the byline said the stories were written by the platform’s “Money Staff,” they were not. Language machines wrote them, but the massive errors (described as moments of hallucinations) exposed the site! Actual staff members saved the day by cleaning up the copies.
• Sports Illustrated’s Street Spin: The Street, a publication from the stable of Sports Illustrated, published a raft of bot-generated stories and made matters worse by curating and attaching fake author identities to the stories. The spin didn’t end well.

But there have been good experiences, as well:

• When analysing large data caches, modelling for pattern or image recognition machines have proved very valuable. Media houses such as Buzzfeed News and The Wall Street Journal have used AI to establish significant trends in otherwise desperate and solitary occurrences or patterns. Examples range from the Mauritius Leaks, which involved 200k highly technical documents, to the expose on the miles of dangerous lead cables around New Jersey streets that posed severe public health risks to residents. Zach Seward also documents a few other examples in his piece entitled “AI news that’s fit to print.”

Where is the Nigerian storyteller in all this?

Professor Farooq Kperogi and I collaborated on an academic paper for the Journal of Applied Journalism and Media Studies entitled “Light in a Digital Blackhole: Exploration of Emergent Artificial Intelligence Journalism in Nigeria.”

The purpose of the exploratory study was to answer the following questions:

1. What is the state of automated journalism in Nigeria?
2. How has AI changed the business and practice of journalism in the country?
3. What is the attitude of Nigeria’s traditionally technology-averse journalists toward automated journalism?
4. Does the emergence of AI threaten the job security of Nigerian journalists?

For this study, and with the help of Media Edge Polls, we interviewed 14 media executives through WhatsApp, Facebook, and X (formerly Twitter) between October 25 and October 31, 2022.

The study found that social media and the rise of citizen journalists have changed the landscape and accelerated the adoption of automated journalism in the mainstream.

More and more media houses are using tools, including social media integration software like Echobox, Hootsuite, Revive, and Dlvrit, to drive audience and revenue goals. The election watchdog, Yiaga Africa, collaborates with some TV stations to collate and analyse election results using AI tools. Automated fact-checking systems, drones and language management tools are also being deployed.

Which tools can I use?

The choice of tools depends on what you are doing or want to achieve. Today, many storytelling tools and avatars can tell a story with illustrations, charts, graphics design, and even some background music. A few of them can even do it better than journalists. Seven notable ones are:

1. Jasper AI – for generating writing needs.
2. Rytr – for structured plots.
3. Writesonic/Perplexity – for detailed research.
4. Sudowrite – to enhance the writing process for novelists.
5. Plot Factory – for creating story outlines and character names.
6. NovelAI – for immersive writing experiences and character visualisation
7. Chatgpt – a chatbot that uses natural language processing to create human-like conversations.

In a story entitled “After Hell, It’s Lokoja”, for example, LEADERSHIP used a drone to tell the story of the floods that submerged swathes of the town when it was impossible to reach many places there by road.

While costs and infrastructure remain significant barriers to adoption, attitudinal differences between younger journalists and the older, more established ones were also noticed, with newsrooms embracing more diversity in age cohorts and educational backgrounds.

Job losses? What jobs?

Our study did not justify the fear of imminent job losses among Nigerian journalists. However, the impact of the disruption on readership/audiences and revenues due to economic reasons and changing demographics is undeniable.

If anxiety about job losses leads to greater introspection, retooling, and adoption of technologies and practices that improve journalism, especially the core business of storytelling, then it would be a good thing.

There are more of us now than there has ever been – citizen and career journalists alike – with significant resources at our disposal to decide what the news is, what it should mean, and what actions it should spur.

In their book on participatory journalism, Jane B. Singer and her colleagues make a poignant point on the rapid and uncertain pace of change we are experiencing when they say, “Today, journalism is on a journey into uncharted territory – and the road is crowded with all manner of travellers.”

If the destination is uncertain, the least we can hope for is that we are in good company, human or otherwise. And it won’t matter if the chatbot delivers this lecture next year!

Thank you for listening.