2 min

Startup Sanas received $32 million in seed capital, including a contribution from Google. The company’s technology changes the accents of speakers. Investors have high expectations, but abuse is lurking.

Sanas is in its infancy. At this time, the startup is exclusively active in the United States. Its technology, however, has global potential. The concept is simple. You boot up the software, choose an accent and speak. The technology applies the accent to your voice in real-time.

The idea is similar to existing voice modifiers. The difference is that the result is indistinguishable from a real accent. Investors have high expectations, made evident by the most recent investment round shows. Sanas raised €32 million dollars, including a contribution from Google Ventures. Experts estimate the market value at $150 million. Not bad for a company founded two years ago by three students in their twenties.

Good intentions

The founders met during an AI course at Stanford, an American university. A fourth friend interrupted his studies due to a family crisis in Nicaragua, his native country. He started working in a call center, but was constantly insulted because of his accent. That’s how the idea for Sanas came about.

The technology is sold to call centers. The potential is much greater than that, but “call centres are low-hanging fruit”, said co-founder Maxim Serebryakov in an interview with TechCrunch. At the moment, the startup has its hands full with further development, as support is limited to English speakers. Call centers make it possible to bring in money at the same time. Eventually, the company hopes to offer the technology worldwide. For example, via integrations in Teams, Zoom and other large communication solutions.

Major risks

The founding story shows that Sanas’ heart is in the right place. Yet, good intentions aren’t always enough. The technology hides the identity of users. The concept has several similarities with deepfakes, wherein AI technology fakes imagery of persons. Sanas works with audio, but the risks are similar. Fraudsters can use the technology as well –not to prevent discrimination, but to deceive others.

The European Commission is working on several laws to regulate the use of AI. At the moment, regulators focus on face recognition. We expect various AI applications to come under scrutiny over time. Sanas is aware of the movement. The company only provides the technology to trusted organizations, hoping it won’t fall into the wrong hands. The question is whether that is enough.

Tip: Microsoft restricts access to several controversial AI services