AI in Journalism Challenge 2023. Final Report
New York: Open Society Foundations (2024), 60 pp.
"This report describes the Applied AI in Journalism Challenge (AIJC)—a competitive accelerator program intended to prototype pragmatic applications of artificial intelligence (AI) capabilities in mission-driven newsrooms around the world. The program was launched in June 2023 and ran until November 2023. Funding was provided by the Open Society Foundations. The author of this report was the lead consultant for the project. The AIJC program was the first of its kind and therefore was itself a prototype—both of a process for accelerating the pragmatic adoption of AI within newsrooms, and also of an approach to developing investment opportunities around applying AI to news. The program was largely successful, developing a cohort of capable and motivated teams and actively engaging with them as they developed substantial new capabilities. The program also succeeded in producing a range of practical insights for its different stakeholder groups—newsrooms, program operators, investors, and the news industry generally.
This report provides a complete record of the AIJC program. The primary intention in writing it is to contribute the experience and insights accumulated during the program to the global journalism community, and to provide sufficient detail and specificity to be directly useful to anyone seeking to apply those lessons to hands-on projects, accelerator programs, and investments. The first section of the report provides a systematic description of the program, including why it was initiated, what it was trying to achieve, how it was set up, and how it actually played out in practice. The second section reviews the tangible outcomes of the program and describes the “funneling” of projects from the proposals in the initial applications, to the teams selected to participate in the program, to the finalist projects, to the eventual program winner (and runner up). The third section lists the specific lessons learned by the AIJC community, one by one, with a detailed explanation of each lesson and the evidence for it from the program. The fourth section gathers and analyzes these lessons to provide a succinct set of specific recommendations, organized for newsrooms, for operators of similar AI in news accelerator programs, for investors, and finally for the global news ecosystem. The fifth and final section draws on those lessons to look into the near future and offers some general opinions and guidance for changes that we might anticipate for news in the age of AI, as well as some guidance for preparing journalism for that transformation." (Introduction, page 4)
This report provides a complete record of the AIJC program. The primary intention in writing it is to contribute the experience and insights accumulated during the program to the global journalism community, and to provide sufficient detail and specificity to be directly useful to anyone seeking to apply those lessons to hands-on projects, accelerator programs, and investments. The first section of the report provides a systematic description of the program, including why it was initiated, what it was trying to achieve, how it was set up, and how it actually played out in practice. The second section reviews the tangible outcomes of the program and describes the “funneling” of projects from the proposals in the initial applications, to the teams selected to participate in the program, to the finalist projects, to the eventual program winner (and runner up). The third section lists the specific lessons learned by the AIJC community, one by one, with a detailed explanation of each lesson and the evidence for it from the program. The fourth section gathers and analyzes these lessons to provide a succinct set of specific recommendations, organized for newsrooms, for operators of similar AI in news accelerator programs, for investors, and finally for the global news ecosystem. The fifth and final section draws on those lessons to look into the near future and offers some general opinions and guidance for changes that we might anticipate for news in the age of AI, as well as some guidance for preparing journalism for that transformation." (Introduction, page 4)
The Program, 4
OUTCOMES, 15
1. Agência Pública (Impact tracker, audio article reader)
2. Raseef22 (A search engine optimization workflow)
3. The Initium (A Chinese language monitoring tool)
4. The Conversation (“Microsites”and reversioned content)
5. Cuestión Pública (Story context from structured data)
6. Zamaneh Media (Newsletter production)
7. Rappler (New brand using AI reversioned content app)
8. PumaPodcast (Workflow efficiencies for news bulletins)
9. Scrolla (A multimedia workflow using AI reversioning)
10. Rubryka (Contextualizing stories from archives)
11. Daraj (Impact tracker)
12. Meduza (A summarzsation pipeline)
Lessons, 30
Recommendations, 45
Next steps
OUTCOMES, 15
1. Agência Pública (Impact tracker, audio article reader)
2. Raseef22 (A search engine optimization workflow)
3. The Initium (A Chinese language monitoring tool)
4. The Conversation (“Microsites”and reversioned content)
5. Cuestión Pública (Story context from structured data)
6. Zamaneh Media (Newsletter production)
7. Rappler (New brand using AI reversioned content app)
8. PumaPodcast (Workflow efficiencies for news bulletins)
9. Scrolla (A multimedia workflow using AI reversioning)
10. Rubryka (Contextualizing stories from archives)
11. Daraj (Impact tracker)
12. Meduza (A summarzsation pipeline)
Lessons, 30
Recommendations, 45
Next steps