Blog 04

To hear an artificial voice that sounded so similar to mine, while knowing that it was far from my actual articulations, was a strange and even mind boggling experience. These text-to-voice technologies, such as the software developed by Lyrebird, are as fascinating as they are dangerous.

This voice cloning phenomenon poses serious threats to the political world. In a society where “fake news” is so prevalent, technologies that can be the source for forged information can be a consequential danger. The uprise of computerized voices, as talked about in NPR’s article “Can You Believe Your Own Ears?“, makes way for a whole new era of attempting to verify falsified information. With systems like Lyrebird publicly available, it has become increasingly easy to claim that true information is inaccurate and vice versa.

While these technologies pose a great threat to the political world, since it generates a wider range of public information that may not be valid, they are extremely beneficial to the medical field. As talked about in The New York Times’ article, “Helping My Nonverbal Son Find His Voice”, people with diseases such as cerebral palsy can use these text-to-voice programs in order to generate sound, something that they otherwise would never be able to do. This sort of aid is revolutionary, since the toll of losing your ability to have a voice can be detrimental to mental health and development in children.

Joe Nero 

 “And when you’re a star, they let you do it. You can do anything. Grab ‘em by the pussy. You can do anything.” When this recording of then presidential candidate Donald Trump in 2016, America and the world erupted. Many thought Trump’s chances at becoming president were buried solely because of his 2005 comments degrading women. The recording comes from a private conversation with former television personality Billy Bush. While Trump’s career was ultimately not affected negatively because of the tape, Bush lost his job and has a negative tinge on his legacy because of one three minute conversation.

I don’t bring this instance up as an attack against Donald Trump. Rather, the tape exhibits the fact that recordings, videos, pictures, and any other form of digital media spread like wildfire in today’s world, and that what you say or do in front of a camera or microphone can destroy your reputation and possibly even your life. This is where the Lyrebird technology could become extremely lethal in the future. As voice replication technologies grow and improve in the future, discerning one’s true identity will become more sophisticated. American politics, especially in a campaign setting, already thrives off of the degradation of the other candidate(s). Attack ads are commonplace for both sides of every race, many including sound bites from the targeted candidate. With the emergence of voice replication, negative campaigners will become more ruthless in their ad making, and will use the technology to sabotage opposing candidates. Tapes similar to that of Trump will become more and more prevalent, and in a country where the voters already do not trust politicians it will be extremely difficult for candidates to clear their names. Voters will have no real way of knowing what a candidate really said or did. The unfortunate message to those negative campaigners becomes the same that Billy Bush heard from Donald Trump: “You can do anything.” 

The Lyrebird team has pioneered text-to-voice technology, specifically with the creation of custom voices users can make based on their own voice, as well as recreation of existing people’s voices such as celebrities or politicians. The technology has obvious and beneficial applications in the fact that it can help people have a voice to use if for some reason they don’t/have lost the ability to speak themselves. Lyrebird in fact has a project dedicated to this very situation that they call “Project Revoice”. “Project Revoice is a non-profit initiative with the ultimate goal to ensure that no one living with ALS will ever have to suffer being robbed of their voice” (Projectrevoice.org). While projects like this clearly seem like a reason this technology will help society, there are ethical issues that can occur with it. The Lyrebird team themselves has an ethics page where they address how they use the technology but if in the wrong hands others may use this technology in a less innocent manner. The ability to copy voices, particularly of people in power is dangerous for obvious reasons. As covered in the article on risks of Lyrebird, conversations could be faked leading to conflict or other issues. At the same time real conversations that may be incriminating can simply be called fake by the people/person being accused.

 

I was very impressed by the advancements that the Lyrebird team has been able to make with text-to-voice technology. I have been exposed to this technology before due to learning about how Stephen Hawking would use this type of technology to communicate with people. However I was unaware of how easy it could be for anyone to utilize this technology. This software can be very useful for those who have lost their voice, for example those who have ALS or who have suffered a stroke. People with these diseases can use this technology to get their voice back.

Unfortunately text-to-voice technology also has some drawbacks. In the NPR article, “Can You Believe Your Own Ears? With New ‘Fake News’ Tech, Not Necessarily”, the problems with this technology was highlighted. Despite the alleged security of this system, it is still possible for your voice to be used without your permission. If anyone finds recordings of your voice are on the Internet, these recordings can be used and uploaded to Lyrebird to create a virtual avatar of your voice to be utilized for someone else’s purpose. In the article it mentions how influencers could “create fake clips of politicians”. This could be very problematic because then these influencers could cause huge switches in support or opposition for certain politicians. Despite the couple of issues in their technology, Lyrebird is creating a type of software that will be very helpful in helping to spread people’s voices globally even if they have lost their physical voice. It will be exciting to see how this company continues to change the face of technology.

What is most unique about humans is the distinct voice they produce when speaking, a voice from deep within their bodies that no one else can copy. To illustrate, as Cavarero states, “The vibration of a throat of flesh that repeats words with an unrepeatable voice” (522 SSR). It must take a professional actor with great skill to even come close to replicating a voice, although, a new company called Lyrebird has come out with a text to voice generator that is able to copy anyone’s voice and have it say whatever they type in. This raises many red flags and skeptics such as Tim Mark from NPR provides quotes from others who question the safety of this technology, “I don’t think it’s an overstatement to say that it is a potential threat to democracy” (Henry Farid, Chair of computer science at Dartmouth College). The lyrebird team defends their technology through their catalyst to creating the technology, which is a solution to those who have lost their voices to disease. Specifically, Lyrebird has a partnership with the ALS Association, “Project Revoice”, and their goal is to help people with ALS keep their voices and improve their methods of communication. Ethically, Lyrebird also acknowledges the dangers that comes with their technology and states how they will attempt to avoid the misuse of it through global awareness and the customer’s voice cannot be released without consent. Personally, I do not see any reason for this technology to be open to the public, it would be a matter of time before someone figures out how to hack the system for their own benefits. Evidently, I believe Lyrebird should be a private company and only be in partnership with companies such as the ALS Association. The less people this technology is available to, the less chances they are of it going to the wrong hands.

Text-to-voice technology and vocal cloning is a phenomenon that is something that I have never seen before and never knew it was so easily accessible by basically anyone who has access to the internet. Lyrebird has led these innovations and is now one of, if not the, most progressive companies in this field. I feel this technology has both benefits and drawbacks. Text-to-voice and vocal cloning can be used to aid society in a variety of ways, such as those with difficulty speaking (such as Lyrebird’s partnership with people diagnosed with ALS), audiobooks, brand voices, and video games. However, there are also risks that come along with this technology. The first and most obvious would be the ability for others to create audio clips of another person’s voice to say things that they have never and would never say in an attempt to slander them. A very easy example of this would be political elections. Lyrebird also states that the audio clips any person submits to them are completely safe and never shared with anyone. I believe this is a feature that is necessary in all companies that may pursue vocal cloning in the future to ensure the safety of its users. Lyrebird also says that their technology is still in its early stages, which makes me eager to see just how advanced this field will become in the near future as the vocal cloning they have already produced is truly amazing.