Category: Culture

Voice actors, tricked by LOVO into creating AI replicas, file suit.

LOVO, Inc.

A class-action suit was filed last week by voice actors Paul Lehrman and Linnea Sage against AI developer LOVO, Inc. According to the complaint, LOVO induced the actors to provide recorded material under false pretenses—material which was then used to produce synthetic replicas of their voices to become part of a catalog offered to paying customers. The complaint also alleges that LOVO defrauds its customers who believe they are using voices that have been legally obtained.

Both Lehrman and Sage were contacted via the freelance hub FIVVER and solicited for voice work. Both asked about the ultimate uses of the recordings—a standard question which affects the price an actor will charge—and both were lied to, according to the complaint. Lehrman was told that the recordings would be used exclusively for academic research, and Sage was told that hers would be used to produce “test scripts” for radio spots. In fact, both anonymized parties who contacted the actors were employees of LOVO—co-founder Tom Lee allegedly contacted Sage—and the sound recordings were used to train the company’s AI to create replicas of Lehrman and Sage’s voices.

Not only were the actors’ voices added to LOVO’s catalog, but the company also used the replicas for its own marketing and capital-raising purposes, renaming Lerman and Sage as “Kyle Snow” and “Sally Coleman” respectively. For instance, the complaint alleges that Sage’s voice was used in demonstrations to raise millions of dollars in venture capital, and Lehrman discovered LOVO promoting the narration attributes of “Kyle Snow” in its own article entitled “The 5 Best Male Voices for Text to Speech.”

The causes of action in the complaint include violations of New York’s right of publicity (ROP) law, deceptive practices, false advertising, violations of the Lanham Act, unjust enrichment, and fraud. Further, the complaint not only alleges harm to the actors but to LOVO’s customers who are misled into believing that the voices being used in their commercial projects have been licensed or otherwise obtained by permission. Plaintiffs seek millions in damages and legal fees.

Before saying anything else, speaking as a guy who is not very spiritual by nature, LOVO has engaged in the practice of soul stealing. Specifically, for Paul Lehrman to read a description of his performance attributes, having had that talent literally stolen and bottled for sale, is a chilling thought. “With his upbeat tone and slightly faster talking speed, Kyle Snow has the perfect voice for conveying enthusiasm and youthfulness,” the description begins. The prospects of AI replacement in the workforce are problematic enough, but imagine reading your own resume and discovering that it’s actually promoting an AI replica of you made without your permission.

The proceedings of this case may prove instructive to many parties with an interest in public policy related to artificial intelligence. The deceptions, if proven, should be damning to LOVO itself, but this case entails considerations that will be worth watching, even where AI development is conducted without lying to obtain training material. Specifically, many interests are looking at state ROP law as a basis for expanding related protections for all individuals. This complaint cites NYS civil rights law, which “imposes liability on a party for misappropriating an actor’s voice ‘for advertising purposes or for the purposes of trade without … written consent[.]’”

The use of a performer’s likeness for advertising purposes is central to many ROP laws in the states that have such statutes, but this case implies taking a more universal approach to proscribing the reproduction of anyone’s likeness for almost any purpose without permission. In fact, the sci-fi thriller quality of transferring not just the technical sound of a voice, but the personality of that voice, reaches beyond the concept of “likeness” as applied to date. It adds a new layer of meaning to the crime of “identity theft.”

Creative work is always a combination of natural talent and hard work to develop certain skills, but one need not be an actor, or any kind of artist, for the same principles to apply. If identity comprises thought, emotion, likeness, and movement, which of these attributes must be copied—and with what degree of precision—before “soul stealing” occurs? I don’t know the answer to that, but I will very curious to see what precedents are set by a case like this one.

TikTok Exploits a “Target-Rich Environment” for Drug Scams According to DCA Report

TikTok may be the perfect crucible in which to exploit the frailties of negative body image and then breed scammers who con millions from people looking to obtain drugs for weight-loss. According to a report released today by Digital Citizens Alliance (DCA), a joint investigation with the Coalition for a Safer Web found at least sixty operators, several posing as pharmacies or medical professionals, fraudulently offering to ship the antidiabetic drugs Ozempic or Mounjaro or the weight-loss drug Wegovy.

From report “Ozempic Scams on TikTok: The only thing likely to get lighter is your wallet.”

“The moment is tailor-made for scammers to take advantage of American consumers,” says DCA executive director Tom Galvin in its press release. “An estimated one in six people say they take Ozempic or other weight loss drugs and just as many are considering it. That’s a target-rich environment for criminals and other bad actors. It’s alarming that TikTok allows these scammers to operate so freely.”

Big Tech critic Tristan Harris has compared the Chinese version of TikTok to “spinach” and the version used by the rest of the world to “opium” because the safeguards deployed on the former do not exist on the latter. Harris did not literally mean drug pushing with that analogy and was instead referring to the addictive nature of the platform. But it is little surprise that the TikTok algorithm is used to detect users’ interest in weight loss and then bombard them with promises to deliver name-brand drugs without a prescription.

DCA Report Lays Out the Anatomy of the Scam

First, there’s the offer of a month’s supply, usually for about $200-$400. Next the seller insists upon using cryptocurrency or payment apps like Venmo, Zelle, or PayPal and to process transactions as “friends and family” to circumvent refund mechanisms. In some instances, the scammer will inform the customer of a “holdup in customs,” which can be expedited by a one-time payment. And finally, the customer is asked for a screenshot as proof of payment, which then provides the scammer with information that can be used to directly trigger fraudulent transactions. As the report states:

A day after making a payment to these scammers, one of the investigator’s credit cards was compromised. Nearly $2,600 was charged to Hertz Rental Car. In addition, the investigator received a Zelle fraud alert within hours of a purchase.

Naturally, these pharma-cons move around the platform by creating and shedding multiple accounts and identities to evade what little scrutiny TikTok employs to crack down on these activities. This is, of course, familiar territory. In 2011, Google paid a half-billion-dollar fine to the DOJ for ad revenues it received from rogue pharmacies in Canada. In that case, for better or worse, the pharmacies and drugs were often real, but the transactions were illegal. In the TikTok examples, with so many red flag indicators of a scam, one might think that few consumers would fall for these “offers,” but perhaps the intent to misuse a diabetes drug in the first place (a potentially fatal decision) indicates a willful blindness that is every con artist’s dream. The DCA report states:

It can be said that it’s better for Americans to be duped out of their money than to receive drugs – whether Ozempic, opioids, or steroids – that can threaten their health or even their life. But it’s a sad commentary when the “lesser of two evils” is the choice offered to American consumers.

Regardless of consumer awareness, this new report is yet another example of the failed policy of laissez faire when it comes to social media platforms. If the forced sale of TikTok wrests control of the platform from the Chinese Communist Party, that will bring the new owner into the reach of U.S. law, but what U.S. lawmakers then do to protect American citizens is another matter.

Photo by: SIVStockStudio

Site-blocking: can the U.S. finally get it done?

site-blocking

The Motion Picture Association (MPA) has prefaced a renewed interest in site-blocking legislation to combat piracy. Will things be different this time?

When the internet industry killed the antipiracy bills SOPA and PIPA in January 2012, I was a newbie blogger but guessed at the time that those parties had totally blown their wad on that campaign. First, there was the boy-who-cried-wolf limitation suggesting that Google & Co. had deployed too much hyperbole to ever again sound the “death of the internet” alarm. Next, the general belief that “the internet” is inherently a force for good was a notion that waned perceptibly after 2012 and then fell off the cliff circa 2016. Today, neither the general public, the government, nor the press fawns over the “white knights” of Silicon Valley as they did when those bills were scuttled.

None of that addresses the fact that the “Stop SOPA” campaign was a tidal wave of disinformation, but it would be naive to think that facts would win today any more than they did twelve years ago. When new site-blocking proposals begin to make headlines, and the network of tech-funded groups howl BEWARE SOPA 2!!, it will be interesting to see whether the same, or similar, false talking points are effective in an environment that is more skeptical of Big Tech in general.

What is site-blocking and why do It?

Site-blocking today would probably work much as it was intended back in 2011. A complainant would have the burden to prove to a court that a platform is principally engaged in illegal activity (e.g., media piracy) and is operating outside the reach of U.S. law enforcement. With sufficient evidence, the complaining party(ies) would obtain an injunction to deny the platform access to the U.S. market. The basic mechanisms are not much more complicated than that, though we can expect the same network of “digital rights” groups to sharpen the rhetorical pitchforks and again stoke allegations that this process will “break” the internet or that it violates the speech right.

Of course, neither claim is true. Site-blocking is employed as a remedy throughout the democratic world where the internet still functions, and speech rights are not infringed (at least not because of site-blocking). There is no more a speech implication to blocking a criminal web platform from access to the U.S. than there would be by interdicting a cargo ship full of counterfeit electronics. On that subject, the need for site-blocking legislation today is more urgent than it was in 2011, and not just for movies and music.

Although the MPA et al. will naturally focus on sites illegally hosting and/or streaming pirated entertainment, establishing a broader rationale for site-blocking—i.e., getting past unfounded ideological opposition—will serve other business and private interests. Online predators of every type have continued to adapt since 2012; evidence shows that media piracy is integrated with a broad spectrum of cybercrime; and the U.S. lags behind the EU et al. in adopting this basic mechanism of protection.

For instance, small-business owners making creative products sold on eCommerce platforms lack the resources to combat, or the margins to absorb, the pace of counterfeiting by foreign actors. Advancements in small-batch production methods and drop-shipping offer new flexibility for counterfeiters to flood the U.S. market with cheap knockoffs, harming both legitimate producers and consumers. Meanwhile, media pirate sites are delivery platforms for malware used for cyber extortion (including sextortion), identity theft, and direct theft of private and sensitive material from personal and business networks.

So, although the MPA will likely be the most prominent advocate of site-blocking legislation, there are many disparate parties—from small-business owners to advocates fighting online sexual abuse—who may see the value in the U.S. finally adopting a remedy the EU et al. have had in place for a decade or more.

An Interesting Moment

In 2011, it was easy to spread the message that site-blocking was only about “Hollywood” protecting its wealth to the detriment of speech on the internet. It wasn’t true then, of course, but it will be interesting to see whether some form of the same rhetoric will gain traction in the coming years. Specifically, a whole generation has grown to young adulthood since then—kids who never heard the proverbial boy cry wolf the first time. Notably, Torrentfreak reports that GenZ exhibits a high rate of pirate site access, citing familiar rationales that streaming subscriptions are too expensive and/or that interest in one title militates against subscribing to the necessary channel.

But what will really be interesting to watch over the next few months will be GenZ’s susceptibility (or not) to the “Save TikTok” campaign already underway. On April 24, President Biden signed a solidly bipartisan law stating that TikTok will be banned in the U.S. unless, within nine months, owner Bytedance sells the platform and, thereby, severs all ties to the Chinese Communist Party (CCP). Bytedance, in addition to vowing it will fight the law in U.S. courts, has already launched a PR campaign, including social media messages that will tap into the same emotional triggers used during the “Stop SOPA” campaign.

As Google & Co. did in 2011/12, Bytedance will use its addictive interface to promote the message that its business interests are synonymous with their users’ rights, only this time, the rhetoric isn’t coming from Big Tech filtered through the Electronic Frontier Foundation—it’s a psyop of the CCP. The efficacy of the “Save TikTok” campaign will be telling, not only about the viability of site-blocking legislation, but about the hoped-for savvy that “digital natives” might reveal about navigating the perils of cyberspace.

In 2011, it was frightening to watch the platforms use the insidious power of the platforms to advocate the policy interests of the platforms. Now, that same playbook is being run by a foreign adversary targeting 170 million 18-29-year-olds, and it is an anxious moment, to say the least, waiting to see how they respond. Regardless, the underlying rationale for site-blocking is sound, and I hope that both copyright and non-copyright interests see it as a necessary protection of American enterprise and security.

Photo by: tommoh29