Media Literacy Education in the Age of Machine Learning

Teemu Valtonen University of Eastern Finland, 
Matti Tedre University of Eastern Finland, 
Kati Mäkitalo University of Oulu, Finland
​Henriikka Vartiainen University of Eastern Finland, Finland
Overview: 
With recent changes in media, where information is shared like a virus, “fake news” resurfaced as a more powerful adversary. One of the reasons “fake news” is so pervasive is that it has become easier to make it look more credible and dynamic. “Fake news” is designed to be sensational, interesting and believable, it often includes real facts. Machine learning algorithms aid in dissemination of “fake news” and polarization of opinions.  “Coping with the world of algorithmically created and distributed news is a challenge to existing media literacy education” Media literacy needs to emphasize individuals’ active inquiry and critical thinking, people’s skills, and ability to reflect and analyze media input.
The article also discusses the algorithms that run online media platforms. They are proprietary to the companies who own the social media platforms, they are well hidden and get a user their personal media feed that is tailored to their personal preferences. They are using “attention engineering” to manipulate people’s emotions without their awareness.
Here is a list of Technologies and Educational themes and content I’ve found to be important to share: 
 
Tracking 
How are people tracked in the physical world (phones, health devices, instagram photos with geolocation data, credit cards, utilities, security surveillance, etc.) as well as in virtual worlds (twitter posts, facebook face recognition, loyalty cards, software updates, calendar events, etc.)—and how these are combined to make an extremely accurate profile of you, the consumer.

Recommenders and optimization
How do recommender systems use tracking data from massive numbers of people to very accurately predict what you might be interested in, based on people with similar interests, as well as on content similarity—and how is that information used for marketing, political and social purposes, provision of instant gratification, and facilitating addictive behavior.

Dynamic content generation
How is natural language generation used, for example, to create large farms of “followers” in social media, to create news and fake news that appeal to users’ interests, to design bots that are able to carry out conversations, and to automate journalism. 

Deep learning
How do computers learn to do tasks by looking at large numbers of examples; tasks such as translate text between languages, spot fraudulent credit card transactions, paint or compose music in specific styles, and recognize faces, speech, or objects in photos. 

Reinforcement learning
How do computers learn, by themselves, to take the right actions in complex, unstructured environments, such as keep up a conversation, drive a car, land a plane, learn to play games better than people do, find optimal settings for media content and customer retention, and optimize logistical operation

Attention engineering
How are social media adapting and tailoring their content to maximize the time users spend on their services through, for example, amplifying negative emotions and excitement, magnifying polarization, and finding an optimal balance between gratification and unpredictability.

Content filtering/ curation
How do web 2.0 platforms learn to curate content that the user is most likely to click or to filter content that is likely to drive the user away; for instance, search engines and video-sharing websites adapting to the user’s earlier choices, and newspaper comment sections filtering out potentially offensive or objectionable comments.

​
I found this table to be a comprehensive and informative tool for discussion of media literacy in “post-truth” times.  The authors suggest these areas to teach as a part of the ICT skill set of a modern digital citizen. 

​

---