Artificial Intelligence successfully writes a scientific paper about itself

share to other networks share to twitter share to facebook

Artificial intelligence continues to evolve in ways we don’t expect. Case in point, an AI recently wrote a paper about itself and submitted it for publishing. But just how did this technological marvel occur?

New AI leaves researchers in awe

The AI algorithm, called GPT-3, is a well-used text-based AI neural net that's been used in everything from AI-generated text adventures to wine reviews. In a new development, the AI wrote a 500-word essay about itself, using references and citations to make a true academic paper. Seeing the AI use a staggering amount of research to present itself for publishing has surprised a lot of people, including the researcher in charge; Almira Osmanovic Thunström.


Speaking with NewsCop, Thunström says that GPT-3 took two hours to write a 500-word essay about itself. Apparently, Thunström was in awe seeing the AI write about itself and the researcher claims that GPT-3’s essay was so good that it could have been in an actual science publication.

“Academic publishing may have to accommodate a future of AI-driven manuscripts, and the value of a human researcher’s publication records may change if something nonsentient can take credit for some of their work,” says Thunström.

Read More: Artificial Intelligence is creating patents faster than humanly possible, breaking the law

AI gets a say in its publication

If the AI’s work about itself wasn’t impressive enough, GPT-3 also had a say in if it wanted its work published. Funnily enough, the AI fully consented to have its work published and had no objections when it came to a “conflict of interest.” Thunström submitted the academic journal and it already has an editor, so things are moving along.

First authorship is a huge part of academia and Thunström feels that GPT-3 deserves credit for writing for itself. If things go well, GPT-3 could end up being one of the first AIs to write a paper about itself and the Ai’s researcher feels that it deserves full credit, should the journal push through.

It all comes down to how we will value AI in the future: as a partner or as a tool,” Thunström says.