Twitter bot detector is a real thing, scientists say
Posted On July 9, 2021
A real-life bot detector for Twitter that can detect tweets and identify which ones are fake has been created by researchers from MIT and MIT Media Lab.
The researchers created the bot detector because of the volume of tweets sent by Twitter, as well as because of its popularity among the general public.
“The volume of Twitter is huge,” said co-author Matthew Green, a graduate student at MIT Media Labs.
“The bots are probably about as big as the top 10,000 words of a typical tweet.”
The bot detector works by analyzing the content of tweets that are being sent to the bot and automatically determining whether they contain fake content.
Twitter is required to post verification checks every day, and it’s been hard for people to access the verification information.
“We wanted to build something that was fast and cheap and was easy to use,” said Green.
“It’s something that people are already doing, so why not make it easy for people?”
The researchers are working on a prototype that uses a small piece of software called a sensor that’s designed to detect a variety of sensors, including infrared cameras, microphones and accelerometers.
The bot detector uses a pair of infrared cameras to record the positions and positions of the individual sensors, and the researchers are using the software to detect if the sensors are transmitting real or fake content to the system.
The bot detects whether or not a tweet is real or not by measuring its temperature.
When the temperature of a tweet drops below about 5 degrees Fahrenheit, the bot detects the tweet as fake and it will automatically tell users to “report it as fake” and delete the tweet.
“People tend to take for granted the idea that we’re monitoring their tweets, and we think that’s a mistake,” said lead researcher Jason C. Smith, a postdoctoral researcher in MIT Media’s Department of Electrical Engineering and Computer Science.
“It’s very easy to say, ‘Hey, this is fake.
If you want to see it, you have to click on a link,'” Smith added.
“But it’s very hard to see what’s happening in real time.”
The researchers were able to detect the temperature changes of the infrared sensors as well, and then automatically report that the tweets were fake and delete them.
The team also developed an app that would automatically check whether a tweet was fake and then send a warning to the person that clicked on the link.
The app would also alert the user if they clicked on a fake link and delete it.
This app could be useful for people who do not have a reliable way to monitor their own tweets, but want to check whether or what kind of content they’ve been reading.
“This is a very interesting tool, and I think it will be useful in many contexts, not just for the monitoring of fake content, but also for the reporting of false content,” said Smith.
“One of the things that I like about the research is that it’s open-source,” said study co-founder Michael Geller.
“I can open it up and see what this really does.”
The research was published in the journal Proceedings of the National Academy of Sciences.