The company is exploring a 🖋text watermarking technique that subtly alters word selection to create an invisible identifier in the text. However, OpenAI is taking a cautious approach due to the complexities and potential negative impacts, especially on non-English speakers. Previous efforts to 🔍detect #AI-generated text have struggled with accuracy, leading OpenAI to discontinue its earlier detector. The new watermarking method shows promise but may be vulnerable to circumvention by users.
🖥Blog from OpenAI- here
#ITNews | Source | @Tech_maker