A new approach for regulating social media: AABB

Photo by camilo jimenez on Unsplash

Social media and social media companies have reached over the past years an extreme power within our economy and society. There are many ideas about how this industry can be regulated, but they mostly only address the anti-trust topic or fake news issue. In the following, I have developed my own framework based on a bill of Joshua Hawley (SMART act).

I call my idea AABB. It stands for Addictive Application Behaviour Ban and shall apply to any B2C application (maybe for the purpose of memorability, we can also call it 2A2B). The idea is to forbid specific techniques of cognitive manipulation to keep users attached to their platform. A/B testing nowadays gain popularity by trying to reverse engineer the human mind to spot the weaknesses of each and everyone’s habits and behaviour.

AABB addresses 4 different aspects mainly taking advantage of their users: Content, Notifications, Self-Awareness and Explainability. These aspects enhance the ideas of the SMART act.

In the following, I’ll briefly explain every aspect and propose a set of rules to achieve a fair regularisation to protect the end-users.

Content

Everyone is familiar with the techniques of content creation and presentation. My idea of regulating this aspect targets the infinite scroll. The infinite scroll is a simple method of pretending there is exciting news to the user by continuously loading new content while scrolling down. But also when the feed is being reloaded by a swipe downwards, the feed looks completely different. Sometimes, this happens by accident, and we as a user cannot find the content we just looked seconds before. It appears so trivial, but it is a main keeper of users.

Therefore, the AABB framework proposes the following rules:

1. The infinite scroll is only allowed once per active session for a max duration of three minutes.
2. Refreshing the feed only allows new content to be added before the original order. The pre-existing order of the content must remain the same after the new content.
3. Navigating into a detail page and back must not change the order of the content of the feed.
4. Videos and other moving images are not allowed to be autoplay unless there is an intended action for continuous playing requested by the user (e.g. YouTube playlist).

Notifications

When it comes towards content and content proposal, notifications are essential to raising awareness of the users. There are a variety of different notification types. These can be either trending new content, informing about posts related to a user (like being tagged), proposing possible new interests (like new friends) or direct communication (PM or group chats).

The issue with notifications is the lack of control for a user. We simply cannot easily deactivate a particular type of notifications without getting into a messy management UI. Furthermore, most notifications which appear as push notifications are only telling there is new content. Example: When someone tags you in a photo, you receive a notification with the information, you have been tagged. If you want to see the picture, you must go into the app, which automatically attaches the user into the app.

Against these manipulative techniques, AABB proposes the following:

1. Users must be able to easily manage their notifications on a very granular level.
2. Periodical notifications shall be prohibited unless they are security-relevant.
3. Push notifications must not only be descriptive about an event but contain the content and allowing a short interaction without the need to open the app.
4. Notification types shall be able to be disabled straight by a (push) notification of this specific type.

Self-Awareness

When it comes about handling and regulating social media, a widespread argument is that users shall be able to know by themselves. And this is very true since everyone is responsible for themselves. Nevertheless, when we interact with social media, these addictive tricks are so smart, users aren’t even aware of these.

Hence it shall be mandatory for a user to know not only how he/she interacts with the provider but also how the provider analyses the user’s behaviour. This might sound wordy or complicated but let me outline the proposal of AABB for this issue:

1. Providers must show the amount of time a user spends on their platform in an easy way, grouped into different kind of activities (e.g. interacting in a group chat, scrolling through the feed, posting oneself…)
2. Networks must have a default time limit of 30 minutes a day, changeable by the user.
3. Exceeding a limit shall not easily be bypassed by ignoring a limit alert.
4. Providers must declare what information they collect on which activity (e.g. when a user scrolls through the feed, they collect the type of content, the interaction, the time, the scrolling pace etc.)

Explainability

Followed by the last proposal of self-awareness, there is a significant lack of explainability. This issue is deeply related to underlying technologies like Neural Nets or any other Deep AI technologies. But the field of explainable AI emerges, and it shall be for user protection matters to be promoted.

It appears maybe unclear why “Explainability” is so important, but it’s precisely this, which we as a user do not get. We just interact with a black box. When we like chocolate, it is likely we’ll see a chocolate ad. If we searched for cars, we’d see ads about cars. But where does it stop? What does even “recommended to you” or “based on what you liked” mean? Like based on what is actually? Explainability needs to get into the deepest granular level as possible.

Based on this vision, AABB should propose the following rules:

1. Every content decision taken by any recommender system shall inform the user about key influencers within the decision (e.g. proposed BMW ad results from the user’s interest in cars, interest in sportive status symbols and the matching age group).
2. Explainability must appear at any level of content, they haven’t explicitly subscribed to, no matter if a regular post or ad.
3. Users must have the option to deactivate decision influencers (e.g. the like of a specific product or their age) for content proposal.
4. An overview page of key data relevant for content proposal must be presentable to the user at any time.

You might think some of these proposals are so trivial, right? Yes, many of them actually are. But these powerful mini-features are just an enormous compound of addictive and manipulating techniques hidden in a user-centric design aiming to maximise the time users spent on platforms. More time means more ads and consequently also more money.

I hope you enjoy my AABB proposal. Feel free to comment. Do you think it goes far enough or is it too far, is it preventing innovation or the right step…? Feel free to share your opinion on this topic.

And let’s hope, we’ll see some real action of politics soon.