OpenAI has quietly acquired Weights.gg, a small startup known for voice-cloning technology.
The startup had gained attention among general users for its tools that allowed people to create and share voice models.
According to the source article, the deal happened quietly earlier this year. Weights.gg also shut down its consumer-facing service around March.
OpenAI brings the team and IP inside
Weights.gg reportedly had around 6 skilled engineers.
The startup had also raised about $4 million, or around ฿130 million, through venture capital funding.
After the acquisition, OpenAI brought the engineering team and intellectual property into the company.
The team is now expected to work across different development groups inside OpenAI.
A voice model sharing platform
The article describes Weights.gg as a platform that worked almost like a social network for user-created voice models.
Its public library reportedly included many imitation voice models based on famous people and characters.
Examples mentioned in the source include Taylor Swift, Kanye West, Samuel L. Jackson, Blackpink members, Bugs Bunny, Donald Trump, and Joseph R. Biden Jr.
That made the platform technically impressive, but also highly sensitive.
Voice cloning creates serious risks
The startup’s technology could create convincing cloned voices using only short audio samples.
That ability raised safety and copyright concerns.
Voice cloning can support creative tools, but it can also enable deepfakes, impersonation, fraud, and misinformation.
The article notes that OpenAI has already been careful about releasing powerful voice-cloning tools to the public.
ChatGPT voice features may improve
Industry observers believe OpenAI does not plan to bring back Weights.gg as a public consumer voice-cloning service.
Instead, the source article says the real goal may involve using the technology and knowledge inside existing products.
That could make ChatGPT voice interaction smoother, faster, and more natural in the future.
Safety will matter more than hype
The move also shows how major AI companies continue acquiring specialized teams.
Voice AI is becoming more important, but safety systems must keep up.
The article points to future needs such as digital watermarking, identity verification, and stronger detection tools for synthetic voices.
OpenAI getting stronger voice technology sounds exciting, but also dangerous if handled carelessly. Better ChatGPT voice interaction is useful, but the real win will be making AI voices safer before fake audio becomes even harder to spot.
 Origin: Letsdatascience
