Tour Rules About

Idea

A bot that goes around social media and translates/reduces people's posts to concepts and principles

Image credit: Thewitehatter

Darko Savic

Darko Savic Apr 03, 2022

Please leave the feedback on this idea

Originality

Is it original or innovative?

Feasibility

Is it feasible?

Necessity

Is it targeting an unsolved problem?

Conciseness

Is it concisely described?

Bounty for the best solution

Provide a bounty for the best solution

Bounties attract serious brainpower to the challenge.

Social media bot that reduces people's posts to principles and/or high concept summaries. It considers both what people say and what they imply. It submits the summary as a comment under the original post.
Why?
Save readers' time by translating peoples' posts to high signal, low noise.
The agenda behind peoples' words isn't always obvious. The bot brings this aspect to light.
Help people put their thoughts into more efficient words.
Make people more honest on social media. You never know when the bot can show up and expose the agenda behind your words.
Make people more conscious of the fact that their inefficient writing might be wasting other people's time.
A tool for checking your own posts to see if you have expressed yourself clearly.
How it works
The bot takes the original poster's words and tries to reduce their meaning to the underlying principles or concepts. At times this could be useful, at other times it could be funny.
In places such as Reddit, the bot could just go around and comment on people's posts without being asked to do so. On other platforms such as Twitter, the bot would only participate when someone asks it to.
The engine underneath
A successor of GPT3 several generations down the line will likely be required for this software to work well. It would probably work with GPT3 right now, but poorly.
In addition, a deep learning algorithm would comb through hundreds of thousands of examples alongside a big team of psychologists/sociologists.
The learning process
Experts would review endless streams of posts from Twitter, Facebook, and Reddit. The team members would pick a random post and reduce it to principles, concepts, or otherwise summarize it to reflect what the original poster was trying to convey. After one reviewer was done, the 2nd would look at the same post and write their own thoughts. Then the 3rd.
A fourth reviewer would compare all 3 reviews with the original post and decide whether they have a consensus. If they do, the deep learning algo is given the data to learn from. If there is no consensus, the data is also provided to the algo but shown as an example of what NOT to do.
But, but.. how does the bot know what the original poster really meant?
The bot knows what other readers know. In other words, if someone wasn't clear enough for the bot, they weren't clear enough for other people as well. If the algorithm is not (yet) good enough to understand what people do, it should be upgraded. If the message was not clear enough, the original poster needs to set people straight. The bot interpreting the OP's words creates an opportunity for the OP to elaborate on what/how they really meant.

artificial intelligence social media communication

2

Creative contributions

Summarizing and collapsing posts into tabs

Shubhankar Kulkarni

Shubhankar Kulkarni Apr 08, 2022

The idea is great. I see two problems with it:
How to summarize, as J. Nikola menitoned?
After the comments are summarized, there will be numerous comments that mean the same thing adding to the redundancy.
Therefore, the bot can collapse similar posts. For example, if people are aggravated by the original post and show their disapproval via the comments, when the bot reads every comment, it understands the pattern. It sorts the comments by the emotion in it. All comments of disapproval fall under a "disapproval" tab. All positive comments collapse into an "approval" tab. The tab (just like the emoji tab) tells you how many comments are in favor and how many are against the post. This way, all the posts are summarized and the response of the entire population (that responded) is summarized too. By viewing the post, a reader will quickly understand that 5k people approve and 500 disapprove.
If you want it to be more specific, the bot could resolve the comments further. It could sort the comments based on the reason the user provides for their disapproval. Under the disapproval tab, there will appear a multiple-choice question for the new reader. The multiple choices are the ones suggested by previous commenters on the post. The new user can then select the reason for disapproval instead of typing the same thing. This will further reduce the bot's work since more and more people will simply choose their option and move forward. If someone has a different opinion, they can always use the text field to type in their comment. The new comment will be summarized by the bot and sorted accordingly, or it may create a new tab/ choice if the thought behind the comment is new.
There can be a toggle view for each user. They could view the post as "sorted" or "chronological", which is the currently used view.

Please leave the feedback on this idea

jnikola

jnikola2 years ago

What if someone comments that the idea is great, but they see some serious drawbacks? Would you count that as approval or disapproval?
I like the idea of the multiple-choice questions for the new readers! It could really work. That's also something you could implement in a similar way on this platform. Imagine people contributing to the idea of being able to see just the titles of the previous contributions, allowing them to easily scan what has already been said. I see it as an expandable title each in a new row or a set of "cards" (similar to the current home page view) with a title and a two-sentence summary. Darko Savic 

Please leave the feedback on this idea

Shubhankar Kulkarni

Shubhankar Kulkarni2 years ago

J. Nikola I would count that as approval. The general idea of "approval" overlaps with that of an upvote. You can upvote and still suggest improvements to an idea. Similarly, you may like (approve, here) an idea but suggest things that could make it better. "Disapproved" comments are like downvotes, but since things could be subjective, a "disapproval" may not necessarily mean that the idea is disliked in general.
The second part of your comment is interesting and you have my upvote.

Please leave the feedback on this idea

A direct competitor of online writing assistants and what can we learn from them

jnikola

jnikola Apr 04, 2022

The benefits I want to highlight
The idea is great because of many things.
First, I love the idea of having an additional tool that can shape my writing style and guide me to writing perfection (a tool that understands what I want to say)
The algorithm would be great to make the internet a more "drinkable" and understandable place
The mentioned bot would become the direct competitor of online writing assistants such as Grammarly, or more complex ProWritingAid, Frase, Outwrite, Writer (GPT3 based AI), etc.
The problem of implementation
Although similar solutions exist, I like that you wanted to create bot that would roam freely and shape the unclear or too-long posts into meaningful summaries, concepts, and ideas. What I am interested is how did you plan to solve the issue of publishing those summaries?
Possible solutions
Since you would not be able to log in as every user, the bot cannot reshape the original post. It could comment a summarized concept on every post, but that doesn't feel right. Another option is to use it as a Grammarly -an add-on tool for the browser. That way, when your add-on id ON, it reshapes the highlighted text and presents a summarized concept/idea. The last feel like the most possible one.
The problem of understanding
When I did a research on writing assistants for one Innovation Cup, I read that an AI is human-created tool. Therefore, whatever humans can understand, the AI can, too. On the other hand, whatever is not understandable or unclear to humans, will probably be the same for the AI. How could we solve this problem? The key feature could be the context or "knowledge" of the bots who roams around and reads/learns. If the bot is "well educated", it could understand someone's idea better than the regular reader. 
Solution 1
Since there is a lot of data available online, the bots should be divided by fields (fashion, physics, etc.). I think this could increase the summary precision and specificity.

Please leave the feedback on this idea

Add your creative contribution

General comments

Spook Louw

Spook Louw2 years ago

This would also be a good tool for checking your own posts to see if you have expressed yourself correctly. Especially for people who work in English but might think in their own language, it can be hard to be sure that your message is being delivered as you intended. 

Please leave the feedback on this idea

Darko Savic

Darko Savic2 years ago

Spook Louw Agreed, I'm adding this to the "why" section

Please leave the feedback on this idea

Goran Radanovic

Goran Radanovic2 years ago

Could be useful. Instead of five people commenting, "What do you mean?", the bot enlightens them. The only setback I predict with this is the bot construing tone and jokes. If it interprets a message incorrectly, some people may be confused about who to believe—the bot or the OP. 

Please leave the feedback on this idea

ByDarko Savic

AI-powered "what you should know" distractions

ByDarko Savic

AI-generated articles comprised only of quotations from scientific papers and based on what you want to convey

Byjnikola

Teach short courses to people on public transportation

ByDarjan Hren

Influence emotions through visual meditation

ByDarko Savic

Using big data to match compatible people and form great friendships

ByDarko Savic

Your to-do list as social media where only your friends can add activities and vote to prioritize them

ByDaami A.

What If You Had Your Own Personal Stylist?

ByDarko Savic

Personal omnipresent content recommendation algorithm