OpenAI is limiting the wider distribution of its Voice Engine technology as it assesses potential risks, given concerns that, akin to image and video creation tools, it could be misused to disseminate disinformation or enable criminal impersonations online and in phone communications. (Source: Image by RR)

OpenAI Tests Voice Engine with Early Adopters Amid Safety Concerns

OpenAI has introduced a groundbreaking AI system named Voice Engine, which can replicate a person’s voice from a brief recording, marking a significant advancement in AI technology. This system allows for the creation of synthetic voices that can read text in a voice mirroring the user’s, even in languages different from the user’s native one. As reported in nytimes.com, the innovation comes on the heels of OpenAI’s earlier developments in digital image and motion video generation, further expanding the startup’s portfolio of generative AI technologies. Initially, Voice Engine is being tested by a select group of businesses, reflecting OpenAI’s cautious approach to distributing a technology with profound implications.

The startup is proceeding carefully with the broader release of Voice Engine, recognizing the technology’s potential for misuse, such as spreading disinformation or impersonating individuals online and via phone. OpenAI is particularly concerned about the risk of synthetic voices bypassing security measures in voice authenticators for online accounts, highlighting the need for responsible deployment. To mitigate these risks, OpenAI is considering safeguards like watermarking synthetic voices and restricting the use of public figures’ voices, indicating the startup’s commitment to ethical AI development and deployment.

Voice Engine’s unveiling aligns with a broader trend in AI, where companies like Google and various startups have also ventured into synthetic voice generation. These technologies find applications across various domains, from generating audiobooks to animating online chatbots and even creating virtual radio DJs. OpenAI itself has previously utilized similar technologies for ChatGPT’s spoken version and other business applications, showcasing the startup’s longstanding involvement in the synthetic voice domain. However, the unique capability of Voice Engine to accurately replicate voices from short recordings introduces new possibilities and challenges, emphasizing the technology’s transformative potential and associated risks.

Despite the concerns, Voice Engine represents a hopeful advancement for individuals who’ve lost their voices due to illness or accidents, illustrating the technology’s beneficial applications. OpenAI has showcased the technology’s capability to restore a woman’s voice affected by brain cancer, using a past recording of her voice, highlighting the positive impact Voice Engine can have on people’s lives. As OpenAI navigates the technology’s future, including its role in potentially sensitive contexts like elections, the startup remains focused on exploring the vast possibilities of AI while prioritizing safety and ethical considerations.

read more at nytimes.com