AI is for everyone at Otavamedia

Otavamedia has been closely monitoring the changes and opportunities that have been brought about by the digitisation of work. Machine learning and rule-based AI have already been in use for some time.

 Our basic principle is that artificial intelligence is for everyone, and everyone is encouraged to experiment with AI to see how it could help them in their own work.

 The Digital Learning Path is a Teams channel that was established in 2019 to provide training and enable personnel to share information, and it has been in use ever since. More than 700 AI-themed training sessions have been held since early 2023, so there is definitely interest in the subject. 

Sharing study path insights

Six Elements of AI study paths were organised at Otavamedia during 2024. The final assignments for the course were presented to colleagues, and attendees discovered many ways in which AI can streamline and speed up our work. More Elements of AI study paths will be organised in 2025.

Otavamedia began describing team-specific AI technology landscapes in autumn 2024. Team-specific AI landscapes help to determine which AI tools each team has at its disposal and whether they know how to use them properly. It also identifies any gaps – whether there are any necessary tools that are not yet being used for one reason or another.

Landscape descriptions can be utilised on at least four levels: competence development and improving personnel’s AI literacy as specified in Article 4 of the EU AI Act, managing AI competencies, and determining whether AI targets can be achieved with these tools and competencies. Technical support teams also get a roadmap of the tools they will need to support.

Team-specific AI landscapes help to determine which AI tools each team has at its disposal and whether they know how to use them properly.

AI’s many applications

In editorial work, AI helps with tasks such as article headings and leads, and social media posts. Whatever you do with AI, you need to tell people about it, and fact-checking is always the responsibility of the human.

Artificial intelligence is now part of the toolkit for magazine ADs, that is, the people who create infographics. For example, resizing an image or illustration by extending the background speeds up their work. Tinka Lindroth,an AD for Naistenmedia, thinks that AI is a good tool for illustration.

“Although they arouse mixed feelings, AI-generated illustrations can provide richer imagery, as AI’s creative process is not limited by existing pictorial matter. We no longer rely on image banks.” 

AI has also enabled new additional services for readers. Since 2023, Tekniikan Maailma’s readers have been able to utilise the magazine’s own AI: Tuuma. It’s a new, experimental service that helps you access information that’s hard to find using Google searches alone.

“ChatGPT didn’t have access to TM’s content, so the editorial team thought we could create our own AI to retrieve answers from our TM content. We wanted to build a two-way communication channel between our readers and editorial team,” says Tekniikan Maailma’s editor-in-chief, Sami Rainisto.

Tuuma has now been in service for a couple of years, and the next step is to give it access to the entire Tekniikan Maailma archive. The AI is currently searching for information within the last 5–6 years of content.

“The question is, what kind of response will the AI come up with if Tuuma combines information from 1960s and 1990s magazines in the same answer? Knowledge has had time to accumulate and change.”

TM still has some work to do in teaching people what kinds of prompts will get good responses from Tuuma. Many people use single-word searches, which do not generate the best results.

“Our vision for the future is a talking Tuuma, that is, an AI that responds to voice searches in a human-like manner,” says Rainisto. 

Whatever you do with AI, you need to tell people about it, and fact-checking is always the responsibility of the human.

Artificial intelligence in daily work

For years now, we have been using both machine learning and rule-based AI in our day-to-day work for consumer customers. Generative AI is often used in coding. It is a fast and industrious tool that can independently produce finished code.Generative AI is also widely used in automated marketing –  to produce marketing texts, for example.

Finland’s Council of Ethics in Advertising issued a warning concerning an advertisement for Koululainen (a magazine aimed at children and young people) in the dating app Tinder (MEN 49/2024). Otavamedia was unfamiliar with the way in which Google Ads AI uses data to decide on and seek suitable target groups. Based on the AI, the target group in this context was the children’s parents. Otavamedia is very sorry about this incident.

The ads were immediately removed and all AI-based advertising for Koululainen magazine has been suspended for the time being. Based on what we have learnt, ads for other products published by our company have now been blocked from appearing in a variety of dating services. We are also in the process of identifying other possible ad locations that should be excluded to avoid similar situations.

Ilkka Torkkel, Director of Customer Relations, Analytics and Research, thinks that generative AI is capable of creating simple marketing texts entirely autonomously, but a human currently checks and, if necessary, corrects the text before publication. 

“For example, AI could produce an instructional text entirely autonomously for internal use in, for example, customer service,” says Torkkel.