OpenAI Takes Defensive Stance on AI Voice Engine
“It’s crucial for people worldwide to comprehend the direction in which this technology is progressing, whether or not we choose to extensively deploy it ourselves,” stated the company on its website last Friday. “This is why we aim to clarify the functioning of the model, how it is utilized for research and education, and the safety measures we have put in place.”
In a recent development, OpenAI introduced its Voice Engine, which utilizes text inputs and brief audio snippets of human voices to generate speech that closely mimics the original speaker. This tool has the capability to produce human-like voices in multiple languages.
OpenAI explained that the decision to offer a preview of the technology, rather than a wide release, was made to enhance societal resilience against the growing threat of increasingly convincing generative models.
As part of their ongoing initiatives, OpenAI is actively phasing out voice-based authentication for bank account access, formulating policies to safeguard individuals’ voices in AI, raising awareness about AI risks, and expediting the development of tools to verify whether content is real or synthetic.
Despite these efforts, concerns regarding the technology persist. President Joe Biden’s AI chief, Bruce Reed, has expressed worries about voice cloning, while The Federal Trade Commission highlighted scammers using AI to enhance their activities through voice cloning tools.
In response to these apprehensions, OpenAI released an updated statement to address the fears. The company emphasized its collaboration with various US and international partners to incorporate feedback as they progress.
Furthermore, OpenAI acknowledged that the upcoming GPT4o model for Voice Engine may introduce new challenges. The company is actively assessing and addressing potential risks, including those related to social psychology, bias, fairness, and misinformation.
The looming question revolves around the implications of widespread release of this technology, and it appears that OpenAI is preparing for potential outcomes. The company has yet to respond to requests for comments from Business Insider.