Voice conversion is a widely useful technology that can expand creative possibilities, expand employment possibilities, and benefit people with speech problems. Only a tiny part of its potential to help us has been realized so far. But there are also harmful potential applications, which are important to anticipate and guard against ahead of time.
Voice conversion can be used to fool people into thinking someone said something they didn't. This is deception. It’s wrong, just as lying is. In particular, it is wrong to defraud people or to create fake news, and voice conversion can make it easier to do a good job at these things.
At Respeecher, we are interested in preventing our technology from being used for deception and, more generally, in limiting the total harm caused by bad uses of all synthetic speech technology.
We restrict usage of our high-quality voice conversion systems to non-deceptive content-creation applications by limiting who we work with and what we allow them to do with our technology. If you want to make fake news or commit fraud, you will not be able to use our technology.
The more general aim of limiting the harm caused by bad uses of synthetic voice technology that is not ours is less fully achievable. Whatever controls we establish over our own technology, highly realistic voice conversion and text-to-speech systems will become more and more widely available over the next few years and will no doubt sometimes be used for bad ends. We hope that Respeecher's technology, being early to market, will help educate the public about what is technically possible and make people less likely to fall for deceptive synthetic speech. We also think that an important role in limiting this harm can be played by gatekeepers of content such as Facebook and YouTube, and we are ready to work with such platforms to detect and prominently label synthetic speech. If you have ideas about how we can make synthetic speech technology as positive and un-diabolical a technology as possible, we'd love to hear from you.