Practice and pedagogy: AI tools in the field, dynamic storytelling in the classroom

Richard Mensah Adonu, Iowa State University

Nausheen Husain, Syracuse University

https://tinyurl.com/gijn23-panel

INVESTIGATIVE JOURNALISM IN THE NEW AGE: EXAMINING THE USE OF AI TOOLS IN INVESTIGATIVE JOURNALISM PRACTICES IN GHANA��

AUTHORS

RICHARD MENSAH ADONU¹, REDEEMER BUATSI², ALBERT JUNIOR NYARKO³, MICHAEL SATAAVIEL YERB JR⁴

1. Iowa State University, USA 2. University of Louisiana at Lafayette, USA 3. Washington State University, USA, 4. Liverpool John Moores University, UK

PHOTO CREDIT: GLOBAL INVESTIGATIVE JOURNALISM NETWORK (GIJN)

INTRODUCTION

• Artificial intelligence (AI) has become a valuable tool for professionals across various industries, including journalism. Journalists have recognized the benefits of AI tools, leading to their widespread adoption to enhance productivity.

​

• Broussard et al. (2019) define AI as a field that emulates human intelligence, encompassing machine learning tools, robots, chatbots, and automated writing software that perform human-like communicative functions.

​

• Investigative Journalism remains the last resort of any nation wishing to remain truthful to its citizens and helps unveil matters deliberately concealed by persons in positions of power or accidentally behind a chaotic mass of facts and circumstances (UNESCO, 2023).

INTRODUCTION

• According to (Anderson & Fort, 2023), AI tools can revolutionize investigative journalism by aiding journalists in uncovering patterns, detecting connections in vast amounts of data, and enhancing fact-checking and source-verification processes.

​

• Literature reviews and digital library searches prove that although AI tools for journalism have developed fast, their adoption and use among African newsrooms and journalists still need clarification. Specifically, the adoption and use of AI tools in investigative journalism practices in Ghana are yet to be well documented.

​

• This research thus seeks to fill this gap by examining how investigative journalists in Ghana use AI tools in their daily work, identify the specific AI tools adopted by investigative journalists in Ghana, and discover challenges, if any, in Ghanaian investigative journalists' adoption and use of AI tools in investigative journalism practices.

RESEARCH QUESTIONS

THEORETICAL FRAMEWORK

• Rogers' Diffusion of Innovations theory (2003) suggests that the adoption and use of innovations, such as AI tools in investigative journalism, can be influenced by various factors .

​

• In the context of this study, the innovation is the use of AI tools in investigative journalism. The characteristics of the AI tools, such as their functionality, ease of use, and perceived benefits, will influence their adoption by investigative journalists in Ghana.

METHODOLOGY

• Research Design

In-depth interviews.

​

• Sample and Sampling Method

A purposive sampling approach was used to select five (5) investigative journalists who had experience using AI tools in their daily work.

​

• Data Collection

A semi-structured interview guide with open-ended questions was used.

​

• Validity and reliability

Member reviews.

Findings

• Profile of Respondents:

4 males and 1 female.

​

• Background of Respondents:

Experienced investigative journalists in Ghana.

​

Findings

​

• AI Tools Used by Investigative Journalists:

Yandex, Otter.ai, Chat GPT, Grammarly, Google reverse image search, Monica, InVID Debunker, Open Camera, and YouTube Video Summary.

​

• Most Prevalent Tools:

Grammarly and Chat GPT.

Findings

​

• What Investigative Journalists Generally Used AI Tools For:
• Improving writing skills.
• Conducting research.
• Text editing.
• Video summarization.
• Fact-checking (image, text, and video verification).
• Transcription.

Findings

Specific uses of AI by Investigative Journalists:

​

Findings

Specific uses of AI by Investigative Journalists:

​

Findings

Specific uses of AI by Investigative Journalists:

​

Findings

Challenges Hindering AI adoption by Investigative Journalists:

• Financial Constraints.
• Usage of Incompatible Devices.
• Lack of Awareness.
• Lack of Competence and Negative notion surrounding AI tools.
• Privacy and Safety Issues.

​

CONCLUSION

• Investigative journalists in Ghana are using AI tools in their daily work, with Grammarly and Chat GPT being the most popular choices.

​

• AI tools offer opportunities for improving writing skills, research, text editing, video summarization, fact-checking, and transcription.

​

• Challenges include financial constraints, device compatibility issues, low awareness, competence gaps, network connectivity problems, and privacy concerns.

​

• Findings align with previous research on AI in journalism.

RECOMMENDATIONS

• Conduct training programs and workshops to raise awareness and promote AI tool usage among journalists.
• Foster collaborations between AI experts and media organizations to facilitate knowledge sharing.
• Aim for affordable pricing of AI software to make it accessible to journalists in developing countries like Ghana.
• Allocate funds specifically for AI adoption in journalism to overcome financial constraints.
• Create a community of practice for journalists to share knowledge and experiences related to AI tools.
• Conduct further quantitative studies covering various categories of journalists to expand understanding of AI adoption in the field.

​

​

The End

Beyond Datawrapper: What dynamic data visualizations in award-winning journalism tell us about data education

Authors:

Nausheen Husain, Syracuse University

Travis Weiland, University of Houston

Anita Sundrani, Northwestern University

nhusain@syr.edu

Thematic analysis, open coding

Six winners of the UF Investigative Data Journalism Award:

> Waves Of Abandonment, Grist and Texas Observer

> The Force Report, New Jersey Advance Media

> Lessons Lost, Milwaukee Journal Sentinel

> Bussed Out, The Guardian

> Sacrifice Zones, ProPublica

nhusain@syr.edu

Data journalism and movement on the page

The ability to connect data visualization creations to the back and forth of a journalistic narrative, which must include human stories along with their context, using actual movement on the page, results in effective and award-winning journalism.

nhusain@syr.edu

Research questions

> What makes the dynamic data visualizations ‘dynamic’?

> How do elements of the project’s visualizations support the main takeaways in the narrative?

> What lessons can be gleaned from the project’s visualizations about data education?

nhusain@syr.edu

Frameworks

> “Conceptual understanding and procedural fluency” of mathematical ideas, National Council of Teachers of Mathematics (2000, 2014, 2023)

> Explanatory data visualizations rather than exploratory: “a general switch in journalism from a focus on news and scoops to background information and the explanation of current trends” (Rinsdorf & Boers, 2016)

nhusain@syr.edu

Findings: four consistent techniques identified

nhusain@syr.edu

Finding: dynamic data viz as nut graf

In three of the five investigations we analyzed, the authors chose to use a dynamic dataviz element to explain the crux of the issue at the top of the page; in two of these stories, we experience this in scrollytelling format, which explains the core issue, even before we see the headline of the investigation.

nhusain@syr.edu

Finding: dynamic data viz as nut graf

https://tinyurl.com/grist-observer

nhusain@syr.edu

Finding: dynamic data viz as nut graf

These scrollytelling choices make a compelling case that the author “already knows what the data has to say,” (Steele & Iliinsky, 2011) and is persuading the reader to listen and be moved by the story. The sentence-by-sentence slow-scrolling helps to set the stage for a deep dive understanding of the nuances of the issue.

nhusain@syr.edu

Finding: dynamic data viz for zooming in and out

In stats ed literature (Konold et al., 2015), there are four different lenses through which people see data:

> pointers to the larger event from which the data came;

> case values that provide information about the value of some attribute for each individual case;

> classifiers that give information about the frequency of cases with particularly attribute value; and

> an aggregate that is perceived as a unity with emergent properties such as shape and center

nhusain@syr.edu

Finding: dynamic data viz for zooming in and out

In Bussed Out, the story uses a dynamic data visualization at the top to map one man’s path by bus from California to Indiana (case value perspective). The plain map then morphs into a choropleth map of the homeless rate per 100,000 in each U.S. state (aggregate perspective).

​

https://tinyurl.com/bussed-out

nhusain@syr.edu

Finding: dynamic data viz for zooming in and out

Dynamic data visualizations can help the author facilitate zooming in and out of different views of the data, but also can help facilitate zooming in and out of different parts of the narrative. The movement involved in dynamic data visualizations provides a quite literal indication that the author is now shifting to a different perspective.

nhusain@syr.edu

Finding: dynamic data viz for transnumeration

Transnumeration is a concept that comes from statistical enquiry, and is “a dynamic process of changing representations to engender understanding” (Wild and Pfannkuch, 1999).

https://tinyurl.com/force-report

nhusain@syr.edu

What does this mean for data education?

> The use of movement with data in order to support the narrative, not just the aesthetic of the dataviz

> The use of slow movement to break findings down into their most basic versions for the purpose of clarity

> Explanatory dataviz over exploratory for investigations: in one 2019 study, it took museum visitors a median time of 53 seconds to glean a correct data finding from a viz about plankton

nhusain@syr.edu

What does this mean for data education?

D’Ignazio and Bhargava called for educators to “break the mythology of the all-knowing data visualization” by calling it “information presentation” (2018).

How can we frame dataviz education as information presentation in classrooms?

nhusain@syr.edu

What does this mean for data education?

> Encouraging students to use simple, maybe self-collected datasets to come up with a hypothesis and compile a series of views and statements to prove or disprove it.

> Creating time in data journalism classrooms for students to work with paper and pen/pencil creatively to zoom in and out of data findings.

> Telling stories with user-friendly tools that require step-by-step, simple explanations of data: StoryMaps, Flourish Stories, even social media can be useful here.

nhusain@syr.edu

What does this mean for data education?

nhusain@syr.edu

https://databasic.io/en/culture/storybook