Ask a friend, Sami Kallinen 8-bit-sheep

In the past few months, what have you been experimenting with and learning from it, Founder at 8-bit-sheep, Sami Kallinen?

I have been thinking about examining AI integration in editorial processes!

Overview: In March of this year, we conducted a five-day workshop at Otava Media as part of the 8-bit-sheep and Friends’ optimointiopas.fi project. This extensive research study was commissioned by the Media Industry Research Foundation of Finland to investigate customer experience and data utilization in the media industry. This specific experiment aimed to explore the potential of incorporating artificial intelligence (AI) into the editorial process to improve content production and support journalists’ workflow. Our diverse team set out to assess the current situation, inspire, prototype, and evaluate possible solutions to address the identified challenge.

Context: During this year’s Mediapäivä event, the concept of incorporating AI, specifically language models, into the editorial process was introduced. The objective was to optimize content creation and make the work of journalists easier. To explore this idea, a team of experienced professionals who were passionate about the topic was assembled. This is not new territory for us, as we have been working with language models since the days of word2vec and universal sentence encoder.

Methodology: Having facilitated human-centred design processes for over a decade and completed more than a hundred of them, we have settled on the design sprint as the go-to tool for addressing complex problems. For this experiment, we employed this approach to tackle the challenge. The team participated in a five-day process that equipped them with brainstorming, prototyping, testing with real users, and integrating user feedback to polish the solution. The result was a functional prototype demonstrating the potential impact of AI on the editorial process.

Learnings:

Findings and Insights: The design sprint has demonstrated the great potential of utilising open-source AI technologies and available, although proprietary, large language models in the editorial process to automate certain tasks. This can greatly reduce the monotonous work for journalists. Moreover, by breaking the development down into smaller stages, the team was able to achieve effective outcomes within a limited timeframe.

In user tests involving five content producers, even some of the more hardened journalists were shocked at the quality of the results produced by the AI. Some even resorted to foul language in their shock at it. Nevertheless, we must not let ourselves be enticed by these models; we must also acknowledge the constraints they have. Although they may be efficient in talking, they are not capable of thinking or knowing, as some folks might expect.

Implications on Customer and Employee Experience: The integration of AI promises to improve the quality of content by reducing errors and providing concise summaries that are supported by data and are accurate. These can be composed to create bigger content pieces. Furthermore, resources and time saved on routine tasks can be allocated to activities requiring human intellect, thus improving employee satisfaction levels.

Data and Customer Experience Improvement: Data-driven enhancements to content quality and production processes could create a more personalized digital experience. The project emphasized producing search-engine-optimized content (as long as this is still relevant) and highlighted the need to better understand the digital audience.

Next Steps: Although the project’s future direction is yet to be determined as far as I know at Otava, there are opportunities for additional development in smaller elements or making larger advancements as an organization.

In conclusion, this experiment provided valuable insights and learnings regarding AI-powered solutions for editorial processes, revealing the significant potential to elevate both customer and employee experiences.

Thank you for sharing these brilliant learnings Sami!