How AI-generated content raises revenues and connects audiences

When foot­ball com­men­ta­tor Alan Smith’s words echoed out, “Mes­si shoots — it’s a goal!” as Argentina’s star play­er slot­ted home a penal­ty against France in Qatar last Decem­ber, fans lis­ten­ing in on Veritone’s YouTube chan­nel may have assumed Smith was react­ing to the match live.

What lis­ten­ers actu­al­ly heard was Alan’s AI-gen­er­at­ed voice clone pro­vid­ing real-time com­men­tary from the live match’s ana­lyt­i­cal data. The project, devel­oped by Stats Per­form and pow­ered by Veritone’s AI tech­nol­o­gy, lets fans lis­ten to live game updates from a pro­fes­sion­al com­men­ta­tor on any device in local lan­guages.

It’s hard to fath­om, but this is just one of the many ways that gen­er­a­tive AI is help­ing media, enter­tain­ment and oth­er indus­tries cre­ate more engag­ing con­tent that can con­nect with audi­ences any­where in the world.

Gen­er­a­tive AI also allows com­pa­nies to more effec­tive­ly scale con­tent to engage increas­ing­ly diverse, multi­na­tion­al audi­ences. For those look­ing to fur­ther mon­e­tise con­tent dis­tri­b­u­tion, it’s a sig­nif­i­cant oppor­tu­ni­ty.

Powering the creative engine

Voice cloning offers sev­er­al appli­ca­tions for con­tent cre­ators. It can be used to nar­rate scenes or cor­rect audio in post-pro­duc­tion with­out the actor com­ing into a stu­dio to record it them­selves. It also has the poten­tial to trans­form the dub­bing and
trans­la­tion indus­tries.

“You can use it for dub­bing in an actor’s unique voice in dif­fer­ent lan­guages,” says Ash­ley Bai­ley, direc­tor of prod­uct mar­ket­ing for AI voice at Veri­tone. “You could have Kevin Costner’s voice, the way he speaks, his inflec­tion, his tone, but in Span­ish, Ital­ian, French or what­ev­er lan­guage to bring a more authen­tic expe­ri­ence to audi­ences.”

The same can be said for pod­casts, where AI can open up new mar­kets by trans­pos­ing a host’s voice into dif­fer­ent lan­guages.

While AI can clone any voice, doing it eth­i­cal­ly is crit­i­cal to the medium’s suc­cess. A Dan­ish pro­duc­tion com­pa­ny recent­ly worked with Veri­tone on a doc­u­men­tary about a real­i­ty TV star that had passed away. They want­ed his voice to nar­rate the pro­gramme, high­light­ing the legal issues around this appli­ca­tion of AI.

“You absolute­ly need to get sign-off from the per­son you’re cloning,” says David Can­dler, senior direc­tor of cus­tomer solu­tions at Veri­tone. “If they are deceased, you need per­mis­sion from their estate and the IP own­er of the voice train­ing data to ensure that it has the rights to repro­duce the voice.”

There are two main ways that AI can gen­er­ate cloned speech: text-to-speech or speech-to-speech. “Using a speech-to-speech mod­el where you speak in, and their voice comes out, you’re cap­tur­ing the into­na­tion, speed, emo­tion. It’s fright­en­ing­ly accu­rate,” says Can­dler.
Voice cloning can even allow com­pa­nies to cre­ate new adver­tise­ments by brand ambas­sadors in mul­ti­ple lan­guages with­out them hav­ing to par­tic­i­pate in addi­tion­al record­ing ses­sions.

“You can tran­scribe a cam­paign, trans­late it into mul­ti­ple lan­guages and cre­ate addi­tion­al ads and assets at a scale that we can’t do on our own as humans,” says Bai­ley. “So, it goes way beyond hav­ing gen­er­a­tive AI help you with ad copy to actu­al­ly cre­at­ing cam­paigns.”

Scaling audience reach and revenue growth

Giv­en the vast con­tent archives that most com­pa­nies pos­sess, AI helps them dis­cov­er and enhance their exist­ing assets to expand both audi­ence acqui­si­tion and rev­enue growth.

Con­tent con­tin­ues to grow expo­nen­tial­ly. More con­tent is uploaded to dig­i­tal plat­forms in 30 days than what the major US TV net­works broad­cast­ed over the pre­vi­ous 30 years. In this cli­mate, organ­i­sa­tions that don’t adopt AI to assist in opti­mis­ing their exten­sive archives could be leav­ing big­ger audi­ences (and mon­ey) behind.

Accord­ing to Accen­ture, near­ly a third of AI pilot ini­tia­tives are scaled beyond their ini­tial scope to deliv­er out­comes across the busi­ness. 42% of those sur­veyed deter­mined that the return on their AI ini­tia­tives exceed­ed their expec­ta­tions last year.

Savvy invest­ments now can bring recur­ring val­ue, giv­ing busi­ness­es more mileage out of their con­tent and stream­lin­ing their oper­a­tions by har­ness­ing AI to replace time-con­sum­ing man­u­al tasks.

“For starters, media com­pa­nies can use AI voice mod­els to pro­duce for­eign lan­guage ver­sions of their back cat­a­logues, help­ing expand their dis­tri­b­u­tion foot­print glob­al­ly,” Can­dler explains. And AI algo­rithms can also help them search through their archives faster, mak­ing it eas­i­er to find moni­ti­s­able con­tent.

“A media and enter­tain­ment busi­ness may have mil­lions of assets that it can poten­tial­ly mon­e­tise, but unless you can find them, you can’t acti­vate them,” says Can­dler. “With Veritone’s AI-based archive licens­ing ser­vice, for exam­ple, we can very quick­ly lift spe­cif­ic moments from vast archives and then sell them to doc­u­men­tary film­mak­ers, agen­cies or net­works. It takes far too long to man­u­al­ly sort through the con­tent to find what you are look­ing for.”

Working side-by-side with human creators

Using gen­er­a­tive AI, a large lan­guage mod­el like Chat­G­PT can be built for a spe­cif­ic brand, which can then deliv­er rel­e­vant con­tent and ideas.

“All of the gen­er­a­tive AI com­po­nents would be very spe­cif­ic to their domain,” says Bai­ley. “It can take all the data of a con­tent rights’ hold­er and pro­vide con­tent rec­om­men­da­tions and cam­paign rec­om­men­da­tions on the back of that.”

Although some con­tent pro­duc­ers may be fear­ful of AI replac­ing them, these inno­va­tions are designed to sup­port humans in their roles. With gen­er­a­tive AI con­tent, humans will still need to go in and edit what has been pro­duced.

“For our voice work, we part­ner very close­ly with the voiceover com­mu­ni­ty. It’s not about tak­ing jobs away–we aug­ment human abil­i­ty,” says Bai­ley. “It helps con­tent cre­ators keep up with con­tent demand, allow­ing them to con­nect with glob­al audi­ences in a more authen­tic, per­son­alised way.”

As peo­ple and machines col­lab­o­rate more effec­tive­ly, gen­er­a­tive AI can cre­ate oper­a­tional effi­cien­cies, improve audi­ence expe­ri­ences, and make new rev­enue streams avail­able at scale.

The mes­sage is clear: AI doesn’t make bet­ter con­tent; it makes con­tent bet­ter.

For more infor­ma­tion vis­it veritone.com/generativeai

When football commentator Alan Smith’s words echoed out, “Messi shoots - it’s a goal!” as Argentina’s star player slotted home a penalty against France in Qatar last December, fans listening in on Veritone’s YouTube channel may have assumed Smith was reacting to the match live.

What listeners actually heard was Alan’s AI-generated voice clone providing real-time commentary from the live match’s analytical data. The project, developed by Stats Perform and powered by Veritone’s AI technology, lets fans listen to live game updates from a professional commentator on any device in local languages.

It’s hard to fathom, but this is just one of the many ways that generative AI is helping media, entertainment and other industries create more engaging content that can connect with audiences anywhere in the world.