OpenAI is reportedly working on a solution to detect whether a text was generated by its ChatGPT model, as some education officials have made the alarming decision to block access to the tool in public schools such as in the US.
The use of AI to complete homework assignments by students has raised concerns among educators about the impact on education. One concern is that language models can be used to plagiarize existing work or enable cheating. In response to these concerns, OpenAI is said to be creating “mitigations” to detect text that is automatically generated by ChatGPT, as reported.
OpenAI has since said that by making ChatGPT available for research preview, it aims to gain insight from its real-world use, which it considers crucial in the development and implementation of advanced, secure AI systems. It is also continuously taking into account feedback and learning from it.
OpenAI has always emphasized the importance of transparency when it comes to the use of AI-generated text. The company’s policies require that users are upfront with their audience when using its API and creative tools. OpenAI is looking forward to working with educators to find useful solutions and ways to help teachers and students benefit from AI.
The ability to discern whether written work was generated by a human or machine will alter the way AI tools are used in academia. Schools would have more efficient ways to enforce bans on AI-generated essays, or they might be more open to accepting papers if they can understand the benefits these tools have to offer students.
OpenAI launched ChatGPT, an updated version of GPT-3, in November last year, but it still has the same shortcomings as all other language models.