UX Designer Creates Tinder Bots To Automate Responses

A User Experience (UX) designer and hacker Robert Winters used AI to create bots that swipe and speak on his behalf to women.

The purpose of the bot is to swipe and chat like a person with women.

The hacker indicated that on dull nights, he likes to spend his time swiping and looking for dates.The massive sink in responses encouraged him to develop an automated response using a bot.

He hosted the site on Github, a software development platform which looks like Facebook’s messenger when opened.

The design serves as Tinder’s web interface, with women’s profiles on the left and the entire conversation on the right of the screen.

The layout also has a terminal window that allows visitors to the site to see what the where the bot is doing.

In the first step of the design, he incorporated a customization feature where users can customize the script for automated conversations, take in the positive and negative sentiments, and then the bot will start a mini conversation.

In the second design process, users can train their bots to learn their preferences.

In the background, an elementary machine learning method will run.

This detects features in pictures say color ie. hair colour, eye colour, facial structures, cheekbones amongst others.

The system learns what you like and makes decisions based on your choices due to machine learning images incorporated into the system.

In the next interview, data scientist Jeffery Li stated that it was difficult for men to get dates on Tinder and other dating apps because their profile pictures were not good enough.

To resolve the issue, he also created an AI that would automate the processes for him.

On a simple level, he used machine learning to make predictions about certain things. The idea behind it was to replicate neurons in the brain to make intelligent predictions based on as much data possible.

He also used deep learning to learn the edges of the faces of different shapes inside an image. In order to do this, he had to extract a lot of data from what he had already used before on Tinder to train the model.

He stated that one problem he encountered was that the algorithm learned just a few because he had disliked most of the profiles. To counter this problem, he used attractive images he found on Google and trained the model with it.

He also developed an automated message for the bot that he could change anytime he got a match or start a conversation.