30 March 2024 – OpenAI, renowned for its innovative AI solutions, has unveiled its latest creation: Voice Engine. This cutting-edge technology replicates human voices, offering a myriad of possibilities and raising concerns about its potential implications.
OpenAI’s Voice Engine follows the footsteps of its predecessors, which allowed users to generate images and videos through simple descriptions. With Voice Engine, a 15-second recording suffices to replicate a person’s voice. Users provide a short recording and accompanying text, enabling the system to produce synthetic voices closely resembling the original.
The versatility of Voice Engine extends beyond language barriers, allowing voices to be replicated in various languages. Despite its promising applications, OpenAI exercises caution in its deployment due to potential risks. Like image and video generators, Voice Engine could facilitate the dissemination of misinformation and enable impersonation, particularly concerning online security measures like voice authentication.
Jeff Harris, Product Manager at OpenAI, acknowledges the sensitivity of the technology, stressing the importance of responsible implementation. To mitigate risks, OpenAI is exploring watermarking techniques and controls to prevent misuse, especially in sensitive contexts like politics and finance.
OpenAI joins other tech giants and startups in advancing synthetic voice technology. While OpenAI has previously deployed voice-enabled versions of ChatGPT and offered voice solutions for businesses, Voice Engine represents a significant leap, raising ethical concerns, especially in the context of upcoming elections.
