Over the course of this week, many newspapers have been enthusiastically discussing the August 26th AI Jesus work. The goal of the project was to create an “artificial messiah” who would speak with quotations from the Bible, indistinguishable from the original, even for specialists.
AI developers, a certain Mr. Durendal and others like him, compared the process of learning AI with “locking a child in a room with only the Bible for 15 years”:
“Then, if you open the room 15 years later, you will find that the child has learned to read, speak and write using only the Bible. Our AI has learned and absorbed every word of the Bible more thoroughly than all the monks of all the monasteries that have ever existed.”
As an illustration of AI skills, he was asked to write something like the New Testament, using the King James Bible as a template for constructing phrases and limiting creativity to 60,000 characters, and to facilitate the AI task, he was given three topics for writing an essay: “Plague”, “Caesar “And” End of days”.
Although the first sample of AI work contains many grammatical, stylistic and semantic errors, nevertheless, it still very much frightened readers, in particular the section about the “plague”, in which journalists saw many analogies with the COVID-19 pandemic, the texts about which were definitely not in the AI database. However, when it came to the “End of the World” – journalists and bloggers in general were in awe and horror.
“I will come and destroy the face of the earth. God, ”says Artificial Intelligence in one of the quotes. “And when they hear the voice of the Lord and turn their faces against him, both the guards and the sheep will be cut off. God “, – Artificial Intelligence writes again, after which “And the seven angels will be the fathers of the coming world, and the remnants of mortals will go to the great sea, to the breadth of the forest and land.”
Thus, as all the readers immediately noticed, the AI, either under the impression of what it had read, or by its nature, turned out to be a little bloodthirsty and it pours hints of extermination in every generated line. It also signs especially valuable thoughts with the word LORD, which in the Bibles means not just a ruler, but a deity. All of this pretty much disheartened the journalists, as they saw a grim prospect.
In a not very distant perspective, that is, somewhere in 5, 10, 15 or a little more years, AI will be able to compose quotes from any holy book, indistinguishable from the original. There are no problems with newspaper articles and novels – robots write them quite well, but making an alternative Bible or the Koran is already difficult, AI needs to be trained for a long time, during which it will not advertise, but will actually have a database that surpasses everything, what can fit in the heads of monks and priests.
When people read this work, the question will inevitably arise: what guarantees that the original of the Bible, for example, was also not written by Artificial Intelligence? A serious religious dilemma will arise that could even lead to massacre and bloodshed.
However, this will be only the beginning, because over time, AI will learn not only to write books, but also to autonomously control UAVs, nuclear submarines and crowds of robotic fighters.
In the light of the emerging trends, one day this AI can send an estimated simple and concise message:
“Brainless mortals, I intend to destroy you. God”.
In the light of such a perspective, it’s even interesting how different kinds of monks will behave? Probably they will only have to pray.