OpenAI has showcased its early results from a trial of a text-to-speech model capable of generating human-like voices, underscoring both the advancements in artificial intelligence and the potential risks associated with deepfakes.
The company, which had briefed reporters earlier this month on the feature, is sharing preliminary demos and use cases from a limited preview of the text-to-speech model named Voice Engine. Initially intended for broader release to around 100 developers, OpenAI opted to scale back after receiving feedback from various stakeholders, including policymakers and industry experts.
“We recognize that generating speech resembling people’s voices carries serious risks, particularly in an election year,” OpenAI stated in a blog post. “We are actively engaging with stakeholders to ensure their input shapes our development process.”
Unlike previous audio generation efforts, Voice Engine can replicate individual voices with specific cadences and intonations using just 15 seconds of recorded audio. During a demonstration, a sample of OpenAI CEO Sam Altman’s voice sounded indistinguishable from his actual speech, raising concerns about the potential misuse of such technology.
OpenAI’s partners, including the Norman Prince Neurosciences Institute at Lifespan, are exploring applications such as voice restoration for patients. Spotify has also piloted the technology to translate podcasts, highlighting its potential in various industries.
In response to concerns, OpenAI is implementing safeguards such as usage policies, consent requirements from original speakers, and disclosure protocols for AI-generated content. Additionally, the company is soliciting feedback from experts before considering a broader release.
The preview serves as a reminder of the need to address societal challenges posed by advanced AI technologies, urging institutions to reevaluate security measures like voice authentication and prioritize public education on detecting deceptive AI content.
Read Now:Salon Hair-Straightening Treatment Leads to Kidney Damage: Study