A Twitterbot that tweets Trump-like tweets has been developed by a postdoctoral researcher in MIT‘s Computer Science and Artificial Intelligence Lab (CSAIL). It uses a deep-learning algorithm to make even more surprising statements than Trump himself.

The deep-learning technique gives the computer the ability to find patterns and essentially learn the way a human brain does. For its creation, the artificial-intelligence algorithm trained on transcripts of Trump’s victory speeches and debate performances, as reported by PCWorld.

On Tuesday night, the Nevada Republican Caucus welcomed Donald Trump as he scored a 46 percent victory. Photo credit: The Week / News Inn

“The algorithm essentially learns an underlying structure from all the data it gets, and then comes up with different combinations of the data that reflect the structure that it was taught,” said its creator, Bradley Hayes.

@DeepDrumpf, named based on John Oliver’s recent segment about Trump’s ancestral name, creates its tweets one letter at a time, so many of them do not make sense at all at first. For example, if the bot randomly begins its tweet with the letter “M”, it is somewhat likely to be followed by and “A” and then a “K” and so on until it types Trump’s campaign slogan “Make America Great Again”.

Hayes was inspired by an existing training model that can simulate Shakespeare, as well as a recent report that analyzed the presidential candidates’ linguistic patterns, which find out that Trump speaks at a fourth-grade level, according to the CSAIL.

Why Trump?

Due to the already mentioned linguistic patterns in Trump’s speeches, Hayes though that his simplistic patterns would make Trump the most manageable candidate to study. But is not the only candidate insight, as Hayes said in a near future a Twitter speech bot could be developed for other candidates like Bernie Sanders and Hillary Clinton, according to CNET.

Much of Hayes’ robotic research right now deals with this types of modeling techniques, he said. He believed this invention would be a good way to learn more about some concepts of deep learning and have a little bit of fun in the process, Hayes added.

Source: MIT Computer Science and Artificial Intelligence Lab