“Will the algorithms define us?”

“No,” said Dr Kim Youngyiil, at the Asia Media Summit in the Maldives.

Trevelling from Korea to the Maldives, on his Kuala Lumpur stopover, he observed that many people were watching shorts on their phones at the airport.

“Wherever you are, whatever your language is, everyone is consuming short form content,” he said.

“There are questions we should be asking to the people who create the algorithms that serve up those shorts. Algorithms are the invisible managers of content that has a lot of influence… they can be good or bad.”

Social media platforms and content delivery companies such as Youtube and Netflix have a lot of power, but they are private companies with a commercial profit motive. They also use their power to influence governments and societies. Youngyiil says we need to know more about how their algorithms operate.

Using an example of the Escape to Korea ‘Better Run’ video, he asked:

“Did you see this tourist campaign with the Squid Game theme on your feed?”

Most workshop participants had not seen it. “That is not a surprise,” he said, explaining that the algorithm is not targeting them. The algorithm is targeting young Squid Game short form video watchers that the tourist bureau wants to attract to visit South Korea.

Different social media platforms have different methods for serving up content to their audience.

Facebook prioritieses social relation.

TikTok is a discovery engine, it doesn’t care who your friends are it just wants to know what you like and play more of it. If you watch 3-5 seconds it will make judgements about you. The system collects lots of information about what you want to watch.

Youtube is more like a search engine – it values long term watching and learns from your in-depth watch habits to recommend more to you. It will also engage with your comments feed.  Some social platforms prioritise content with a good thumbnail, but YouTube highlights Click Through Rate (CTR) above the attractiveness of the thumbnail. “Low CTR comes when people see your thumbnail but don’t click.”

Algorithms have changed the ‘media experience’ into a ‘media interface,’ where humans interact with algorithms to get content.

The old model of media consumption was a one way data flow, but now there is a bidirectional data flow using information gathered from users and AI analysis of content.

The old model:

Sender (the broadcaster) –> Message (content) –> Channel (transmission technology) –> Noise (interference)

As countries developed broadcast regulations they asked questions such as: Who controls the sender? What message is received? Is the message well balanced? Does the message have a real life impact?

The new model is:

User (sender and receiver) <– > Algorithms (mediator) <–> Feedback Loop (mechanism or engine) <–> Content (a unique media environment for each individual)

The questions that regulators and societies now need to be asking are: Who controls the algorithms? How do the algorithms affect the users? Do the algorithms drive the viewing experience in a certain way?

“Recommendation Engines are common, they are data filtering, prediction algorithms that suggest content. But there is a deeper algorithm

“The Narrative Algorithm uses AI to structure, frame and present content as a cohesive story or experience, rather than just a list of items to choose from.

“AI is used to understand what a Drama or Documentary is about, not just from the data that is in the overview. This is the cultural side that has been introduced by AI.”

Netflix and Youtube have 36,000 content categories. No human can interact with all these, but by using AI, viewers can better interface with the content they may wish to see. “To get success with Netflix and YouTube you have to adjust your way of thinking to fit with them and get the best results from their algorithms.”

AI based Story Development software is available to help teams with story creation. He mentioned Celtx and NolanAI, two professional AI apps that help you brainstorm story content. Youngyiil also mentioned NanoBanana as a good tool for for infographics and image generation.

While the latest AI tools are a powerful help for broadcasters, they are not socially aware and may not be culturally sensitive to Asian cultures.

Youngyiil discussed a range of tv shows that are not socially acceptable.  For example, school bullying is a problem in Korea (and elsewhere). The Glory Code is a tv drama that can glorify school bullying.

“AI is serving that up for people to watch, but is it good for society? School violence is a serious problem and we need a serious approach, but Netflix has turned it into an exciting thriller. In  another Korean drama, ‘All of Us are Dead’ the school bullies are depicted as zombies, a fashionable tv theme.

“The Question we should be asking is how can broadcasters ensure content diversity in the age of algorithmic media interface? Can Public Service Media successfully compete for audience attention without compromising its core values?”

He suggested that Asian Public Service Media needs a public service algorithm. Otherwise Asian media will just be part of an international commercial algorithm that does not necessarily prioritise Asian culture and values.

In the same session, Snehasis Sur advocated the case for Public Service Media to bring public service values to new media channels and social media.

“The main feature of what we do is Public Service. In the face of instant reactions on all platforms we should not forget the principle on which we work. One of the important elements of that is accurate news.

“In the social media domain there are various players, some of which promote misinformation and disinformation. Public services is the principle behind our content, if we get it right we will enjoy the trust of the pubic. Audiences still come to PSM when they are looking for the credibility we have built up over many years, we must not lose that as we evolve in the new media environment.

“We need to remember that our audience are ‘citizens,’ not ‘consumers.’ We should not treat them as a commercial product, which other platforms do when they try to monetise or influence their audiences,” he said.

Related reports: