The use of artificial intelligence (AI) writing tools can now be detected by new features that have recently been activated by Turnitin, a top provider of academic integrity solutions. Turnitin Feedback Studio (TFS), TFS with Originality, Turnitin Originality, Turnitin Similarity, Simcheck, Originality Check, and Originality Check+ all offer these features. Before ChatGPT, one of the most sophisticated AI writing tools, was released, the foundation for these features had been laid nearly two years earlier.
The Development of AI Writing Instruments
Turnitin was not surprised by the emergence of AI writing tools like ChatGPT, Google Bard, and others given the diligent research and development of their AI team over the previous few years. “We were all impressed with the progress and promise that GPT-3.5 showed over GPT-3,” said Chris Caren, CEO of Turnitin. Even Elon Musk and other AI experts are calling for a six-month break from creating systems that are more potent than OpenAI’s recently released GPT-4 because they were taken aback by the progress made by GPT-3.5.
New Turnitin Features
By offering actionable and cross-institutional data insights. To enhance decision-making, the new features in Turnitin give educators a strong tool. To maintain academic integrity across their student cohorts. The new features enable educators to quickly assess a submission for the presence of AI-generated text. Give feedback to students in their current Turnitin workflows, claims James Thorley, Regional Vice President of Asia Pacific at Turnitin. More than 2.1 million educators and over 10,700 institutions have access to this capability.
The Problem with Detection in AI
AI chatbots are just one example of the generative AI tools that are expanding rapidly. The chatbot had more than 30 million users. Received about five million visits per day two months after OpenAI released GPT-3.5. Demonstrating the wide adoption of AI writing tools. Turnitin is aware that new AI writing tools are frequently updated, with the claim that they are undetectable. According to reports, the new features on the Turnitin platform can identify AI writing with 98% confidence and less than 1% false-positive rate in a controlled lab setting. But it’s impossible to develop a tool that can reliably identify text produced by AI. The goal of AI language models is to generate natural-sounding text that mimics human-written text. According to Muhammad Abdul-Mageed, a professor who directs research in natural-language processing and machine learning at the University of British Columbia.
The Influence of GPT-4
Turnitin’s new features can currently detect the presence of AI-generated text. But as new AI language models become more potent and adept at producing even more fluent language. They quickly supplant earlier detection toolkits. For instance, the GPT-4 exhibits human-level performance on a variety of professional and academic benchmarks. Despite being less capable than humans in many real-world scenarios. GPT-4 achieved a score in the top 10% of test takers and passed a mock bar exam. The GPT-3.5 score, however, was in the bottom 10%.
Response from Turnitin
Turnitin is dedicated to responding to the subsequent iterations and advancements in AI writing. While placing a priority on student safety as well as the requirements of educators and institutions. So they will keep developing and adapting to meet the challenges posed by new AI writing tools.
The turn-on of new features in Turnitin that can identify AI writing tools is a significant step in preserving academic integrity. Guaranteeing fair evaluation for students. Turnitin’s efforts show a commitment to adjusting and responding to new innovations while prioritising the safety and needs of students and educators. Even as AI language models continue to advance. It’s critical to be watchful and proactive in addressing potential problems with academic integrity as the use of AI in writing increases.