AI-driven feedback chats to enhance the on-platform experience
jnikolaJun 21, 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.
SUMMARY
Feedback through the short chat with personalized AI that remembers what you like, dislike or wrote (sessions, contributions, comments) and asks different questions that help you think, rethink and comment. Additionally, it helps to personalize the content on your home page (personalized feed) and contributes to the rating of the content of the authors.
Problems
Hard to get feedback because it's currently time-consuming and non-rewarding
Current feedback ideas trying to make people give their feedback for cool, but still non-rewarding "benefits"
Feedback tool uniform for different content types/topics (not all ideas or challenges have the same idea of being original)
Feedback used "only" to rate or promote users' content
Would you fill 50 feedbacks just to get access to the newest sessions/ideas (which you were not completely satisfied with the last time) or a few badges on your profile on a platform which is not primarily mimicking other social media and boosting your presence online?
The solution
A feedback in form of a short chat with a personalized AI agent that builds your activity profile and builds feedback summary for the author.
Why?
not all comments or contributions are well thought and are a result of a careful and thorough reading and understanding (result in non-constructive spam-like content)
we need more interesting questions (questions that make you think, connect related sessions, see if data is well written, asks you if it relates to you somehow, ...)
we need better question format (if you like to answer in long sentences, algorithm prefers that; if you like short yes/no questions, algorithm does that)
we need better personalized and professional feedback summary for the author
we need better content personalization
potentiating brainstorming and rethinking about the read content
potentiating content connecting and creative thinking
How would it work?
You make a profile, read contents and start with answering some basic questions. With higher engagement, you start to get more personalized questions and start to think deeply about everything you read. Your feed becomes really interesting and the AI-driven feedback questions become intriguing. All this results in great feedback summaries which are sent to the authors to provide them with professional-level comments and suggestions. Everyone benefits.
Mechanistically, the system would be based on the GPT-3 algorithm developed by the OpenAI. More specifically, these tools or similar custom-made variants would be used: Q&A for the interrogation and data collection, TL;DR summarization for long feedback understanding, ESRB rating for rating the users language use, etc.
Examples:
Scenario 1:
You are the IT innovator great in back-end programming and just read an idea about implementation of a new algorithm in cancer genome research. You cannot comment on the originality, feasibility or necessity, but you think this is interesting because you once read something on implementing these kind of algorithms into programming pipelines, so you start the feedback chat. The AI knows you are a programmer, knows to which sessions and ideas you have commented and asks you if this was understandable. You answer "Yes", and the algorithms asks if you could relate this with something you read before (originality). You answer that you read about the implementation of something similar. The algorithm asks if it was the same program, same use case, etc. You answer shortly, like or dislike the content and the feedback chat ends. Although you are not an expert, AI-driven feedback helped you to understand and rethink about the session and extracted the most out of your answers. These feedback data get simplified and is used to personalize your homepage feed content. Also, it's used to build an AI-driven users' feedback that will eventually be sent to the author of the idea.
Scenario 2:
You are a medical chemist with 10 years experience in pharma companies and uses the platform to find new interesting projects to work on. You read an idea of how to engage users to give better feedback. After the reading, the AI feedback chat asks you if you think this has anythingwith chemistry. You say no. The AI asks how did you get feedback from your coworkers while working in pharma. You write that you mostly talked with them directly, although they didn't always like that. The AI asks if you know any other industry where feedback is really important? Etc.
One more reason why we need it - people are sometimes shy, non-creative or just want to be nice
jnikolaJun 21, 2022
Users' activity on the platform consists of
creation of ideas and challenges,
searching and reading other ideas and challenges, and
interaction with them
Some people will have a bad platform experience due to their shy nature or non-creativeness.
Due to the different skillset of every user, some users engage in all three points, while some just write, read or just interact with other people's sessions. As a consequence, some people who create sessions and comment could have great recommendation results that perfectly match their interests, while some who, for example, just read other's ideas, could get really bad ones. It's discriminative because some users will never be highly engaging or creative.
Some people will never give an honest opinion just because they are nice (or bad).
Liking or disliking the content is not always reflecting the opinion of the reader, since people don't want to be rude or sometimes like other person's content due to topic-unrelated reasons.
By being a sort of a mediator between the user and the author, we could both collect honest opinions and help users to tailor their recommended feed content without public engagement. To give every user an ability to personalize it's feed content (recommended content) and be honest, we should introduce personalized feedback.