The artificial messiah has penned about 60,000 words of prophecy that are eerily reminiscent of the real deal. AI Jesus is the creation of engineer George Davila Durendal, who has dubbed the algorithm a “clone of Jesus”. The artificial intelligence was instructed to read the King James Version of the Bible, learning in the process how to construct convincing scripture-like prophecies.
Mr Durendal likened it to “locking a baby in a room with nothing but a Bible for 15 years.”
He said: “You then open up the room 15 years later and find that its learned to read, speak, and write using nothing but the Bible.”
The engineer then told his AI to pen prophecies on the topics of “The Plague”, “Caesar” and the “End of the Days”.
And since he revealed his creation to the world, AI Jesus has penned more prophecies about “Blood”, “Greeks” and “Wisdom”.
The engineer said: “In these days of trials and tribulations many have turned to religion.
“But what religion is left for those who have averted their gaze from the fables of old to the shiny metal toys of today?
“I present you AI Jesus. An artificial intelligence of my invention created from the King James Bible and nothing else.”
AI Jesus learned from the Bible using a so-called Boltzmannian natural-language processing model.
And the results speak for themselves with passages that could be mistaken at first glance as real New Testament prophecy.
One text reads: “The Plague shall be the fathers in the world; and the same is my people, that he may be more abundant in the mouth of the LORD of hosts.
And although some of the verses do not flow as naturally as real scripture, they demonstrate the power of language-learning AI.
Mr Durendal said: “This orderliness/randomness trade-off is one characteristic of AI language models.
“You can have a more interesting model that takes artistic liberties and produces some glitches.
“Or you can have more mundane, more technically proficient writing. But not both.”
But AI Jesus is not the first time engineers have attempted to create convincing works of art through machine learning.
IBM’s Watson Beat, for instance, used a neural network to create background music for a Red Bull Racing video.
Eric Borgos, the owner of Impulse Communications Inc, similarly used open-source language models to create song lyrics.
The end result is the Song Lyrics Generator, which can be used online.
Mr Borgos said: “I fed it lots of songs to learn from, and eventually, it was able to output lines that sounded coherent.
“Overall the project involved around 50 hours of programming, but the end result was worth it.”