Grammarly has launched Authorship, an innovative tool designed to distinguish human-written text from AI-generated content, with a focus on supporting education and improving document authorship tracking.
Grammarly Unveils Advanced AI Detection Tool to Enhance Document Authorship Tracking
In an era where artificial intelligence (AI) technology is increasingly ubiquitous, distinguishing between text generated by a human and an AI has become a complex challenge. Several solutions aimed at detecting AI-generated text already exist, but with inconsistent results. Grammarly has introduced itself into this arena with a new tool named Grammarly Authorship, which aims to provide more accurate identification of text origins within documents.
Introducing Grammarly Authorship
Dubbed Grammarly Authorship, this advanced detection tool is designed to function across a broad spectrum of platforms. It will be integrated into approximately 500,000 applications and websites, making it a versatile option for users. Unlike other AI detection tools that use algorithms to spot AI content, Authorship uniquely tracks the writing process in real-time. This method enables it to differentiate text typed by humans from that generated by AI or copied from other sources.
Comprehensive Functionalities
Authorship will initially launch in beta within Google Docs for all Grammarly users next month, with expansions to Microsoft Word and Apple’s Pages slated by the end of the year. It will be available in all versions of these applications, including free and paid iterations. Grammarly’s price points include a free basic service, a $12-per-month Premium subscription, a $15-per-month Business plan, and an edition for educational institutions. Authorship will be accessible through all these plans, ensuring broad user availability.
Focus on Education
Grammarly has placed a significant emphasis on the education sector, responding to issues faced by students and educators alike. Jenny Maxwell, Head of Grammarly for Education, highlighted this focus, pointing out the frequent lack of clear AI policies within educational institutions. According to Maxwell, this ambiguity has led to an over-reliance on imperfect AI detection tools, causing confrontations when student work is inaccurately flagged as AI-generated. Authorship aims to facilitate constructive discussions about AI’s role in education and provide transparent insights into the writing process.
Detailed Analysis and Reporting
Once activated, Authorship categorises text within a document based on its origin: human-typed, AI-generated, AI-modified, pasted from known or unknown sources, or edited by tools like Grammarly. Users can access detailed analytics, offering breakdowns on how much of the document was authored by humans versus AI. An overall analysis will reveal writing statistics such as the total time spent and the number of active writing sessions.
One notable feature is the generation of a comprehensive Authorship report. This report includes the full text of the document, with segments colour-coded to indicate their source. An additional replay feature will visually demonstrate how the text was constructed, providing users with a clear understanding of the document’s authorship.
Future Enhancements
Grammarly’s future updates for Authorship will include prompts for students to cite external sources within their work. This feature is expected to launch early next year and aims to further support academic integrity.
Market Context
The market for AI detection tools has been challenging, with notable companies like OpenAI struggling to achieve high accuracy. OpenAI launched and then withdrew its AI detection tool in 2023 due to accuracy issues. Despite such challenges, the development of these tools continues, and the launch of Grammarly Authorship is highly anticipated. Its performance, relative to existing solutions, will become clear once it is deployed next month.
Grammarly Authorship represents a significant step forward in addressing the complexities of text authorship in the age of AI, offering a mix of real-time tracking and detailed post-analysis to better serve the needs of individual users, businesses, and educational institutions.
