At a time when generative artificial intelligence is becoming more and more common, especially in the workplace and academia, there is a growing need to distinguish between artificial intelligence and human-generated content. To this end, Grammarly is launching a new Authorship tool that will allow you to distinguish between handwritten, AI-generated or edited text and copied text.

Grammarly will show users if there is an AI-generated part in their text and provide guidance on how to interpret the results. The new tool won’t show why the text was marked as AI-generated, but it will help users find the parts that look like that and correct them to avoid being falsely accused of using AI and plagiarism.

When Grammarly checks a document for AI-generated text, it will divide it into smaller parts and check them for language patterns that are specific to AI. Users will then receive a percentage score for the amount of content that appears to be AI-generated.

A text can be identified as AI-generated for several reasons. Authors may use language patterns, sentence structure, or other characteristics that are commonly found in generated content, such as repeated phrases or unconventional grammar.

The company advocates the responsible use of AI, so if students want to use generative AI to write a text, the system will prompt them to check the academic integrity policy of the institution and use AI for ideas, not work.

The beta version of Grammarly Authorship will be available to all Google Docs customers starting in September 2024.