Technology
Danish Kapoor
Danish Kapoor

Meta has added Waveforms AI to artificial intelligence -based sound production initiative

Meta develops artificial intelligence sound technologies Waveforms AI He bought his initiative. Established eight months ago, the company was working on projects in which it was difficult to distinguish between human voice and AI production sound.

Waveforms AI gathered $ 40 million in the investment round led by Andreessen Horowitz and reached $ 160 million in this process. Alexis Conneau and Coralie Lemaitre, one of the three founders of the company, moved to Meta after the purchase. Conneau made improvements on the GPT-4O Advanced Voice Mode during his time in OpenaI. Lemaitre worked on Google in the field of advertising strategies.

The projects focused on Waveforms include “Speech Turing Test” and “Emotional General Intelligence”. The first concept measures whether the audience can distinguish the difference between AI and human voice, while the second concept aims to reflect not only the right words, but also the emotional tone. In this way, the company aimed to provide more realistic sound experiences both in interactive assistants and virtual media applications.

Meta accelerates its investments in the field of voice artificial intelligence

This acquisition of Meta is carried out through the new artificial intelligence unit, named Superintelligigence Labs. Last month, the company had purchased Playaı, who also worked in the AI Ses field. Both enterprises will operate under the roof of the same team under the direction of Johan Schalkwyk. Schalkwyk is known as an experienced name previously in the development of speech recognition technologies on Google.

Waveforms AI has closed its own website after the purchase, but the company’s basic vision can still be seen on the LinkedIn page. Here, it is emphasized that AI audio technologies should develop not only technically but also based on emotion and context. It is stated that this approach coincides with the goal of developing the user experience on Augmented reality and virtual reality platforms of Meta.

In any case, the development of such natural and emotional AI sounds can also raise some ethical problems. In particular, measures should be taken on the risks of deepfake audio recordings and identity fraud risks. Meta is expected to strengthen the safety and verification mechanisms while developing this technology.

Danish Kapoor