There’s no shortage of voice and language options available on your sat nav or the apps that get you from A to B, but generative artificial intelligence (AI) and text-to-speech technology have made it even easier for celebrities– and even you– to be the voice that guides you on the roads. That means, if you have the digital nous and the time, you could have anyone from your spouse to Nigel Farage telling you to take the third exit on the roundabout.
But fooling around with these voices comes with its hazards. A study in China found that men (it’s always men) take more risks when their sat nav or traffic app voice is provided by a “sexy” female. In effect, the machismo of male drivers is stimulated by what they perceive to be young, attractive voices talking to them. As a result, they behave accordingly, jumping the lights and taking greater risks to impress these imaginary companions. The study from Fudan University in Shanghai found that men were 40 per cent more likely to jump the yellow light when they were listening to instructions from an “attractive” female voice than with a generic one.
When sat navs first became an indispensable but clunky part of modern life, voice options were limited. During the 2010s the novelty of a celebrity saying “you have reached your destination” and “continue for three miles” reached a brief, if slightly naff, golden age. Options included Darth Vader’s menacing authority, Stephen Fry’s musicality, annoyingSimpsons characters, the voice of KITT inKnight Rider or Brian Blessed shouting “Gordon’s arrived!”
Mr T “pitied the fool” who missed their turning, while Kim Cattrall purring must have been borderline dangerous for some men (probably best not to let the burghers of Shanghai know about her). What effect being directed by the voice of your husband or wife would have depends on how well you get on.
“We’ve done a lot of sat nav recordings in many languages,” says Neil Conrich, the managing director of Voice Squad agency. “These would take weeks to record in total. When you did France or Germany, which are much bigger than the UK, each town and part of each city has to be recorded. Celebrities such as Fry might only record the basics as it would have taken so much money and time to record the same amount with them. That would have been for a flat fee and it would have been fairly large.”
Old favourite TomTom no longer makes celebrity voices available, but neither does the goliath that is Google Maps. If you hanker for an actor or a sports star in your car, the popular app Waze (owned by Google) has plenty of options.
Waze’s “voice pack” favourites include Kevin Hart, Arnold Schwarzenegger, Jeremy Clarkson and, of course, Morgan Freeman. America’s official stand-in for God, Freeman was judged by comparison site Uswitch to have the best voice for navigation– it has a resonant 67hz frequency to help you relax mixed with a high variance range to keep you alert– edging out Liam Neeson and Fry.
With female voices, the Uswitch study, conducted earlier this year, found Cate Blanchett beat, not unsurprisingly, Theresa May. Waze rotates many of its celebrity voices so the one you want may not always be available. Given Freeman’s rate for a commercial is said to be around the $1m mark, it’s safe to say these recordings remain good earners for the rich and famous.
Artificial Intelligence could change that, however. The fast-developing technology could– in theory– provide a limitless range of voices beyond those celebrities who have contributed theirs with permission.
“Even before AI there was declining money in corporate work and now it’s the kind of work that is migrating across into synthesised voices,” says Phillipa Brown, the owner of the Voicecall agency.
“Thirty minutes of audio can give you a reasonably good clone of a voice,” adds Al Black, director of voice-over and translation company Voquent. “We’re already talking about having AI and human voices side by side. Clients want to pay for the voice but [also] have the actor’s voice cloned, so they can edit [the script] without going back for re-recordings. That’s how it’s going to evolve. Any purely human-focused business is going to go down.”
It is possible– should you be feeling particularly bored– to create your own sat nav voice by using text-to-speech engines already available to convert directions into sound files. This includes your own voice, though, as Conrich argues: “Why would you want to do that? You’d bore yourself to death.”
Perhaps even more bizarre– and risky for your marriage– is that you could clone the voice of your significant other and listen to them boss you around even when they are not physically in the car.
“We are very, very, very nervous about AI,” says Conrich. “You can buy some fairly simple software, record yourself and AI will play your voice back pretty naturally. If you’re using it, the costs are minimal. I don’t think it will ever recreate the nuance and timing of actors but we’re worried that for simple work, AI will take over.
“It’s creeping into computer games and where background voices are needed, and we won’t let our artists have anything to do with it. It will cost them their livelihood.”
Two weeks ago the Screen Actors Guild went on strike again in the US after a breakdown in long-running talks over protection from AI for their members working in video games, causing game development to be held up this summer.
So where is this going? You could use AI to generate speech in the voice of your favourite celebrity and then use that to create your own voice pack, though, without permission you’d probably get sued for your trouble.
Alternatively, you could create the voice of a figure who is nolonger with us. Margaret Thatcher perhaps? However, that does not include a U-turn phrase.
“We think about AI every day,” says Brown. “It haunts me. “Having your voice ripped even happens to people who are not really well-known. And it impacts jobbing actors more than Tom Hanks or Scarlett Johansson.”
