AI Briefing: Social platforms start tasting the bitter fruit of generative AI

As social platforms move fast with new AI features, users are already breaking things.

Last week, celebrities including actor Tom Hanks, TV journalist Gayle King and YouTube creator MrBeast all warned of deepfake ad scams featuring their likeness. Meanwhile, Meta’s new AI stickers have already generated controversy as various reports last week showed users easily creating problematic content. And on Friday, the U.K.’s Information Commissioner’s Office (ICO) issued a preliminary enforcement notice to Snap over the platform’s “potential failure” to identify and assess risks to children related to its “My AI” chatbot. 

These problems probably aren’t surprising, especially after experts have spent the past year warning of risks with AI-related misinformation and data privacy risks. However, they show how all the worries are more than just a hypotheses — and provide concrete examples for how the AI era might accelerate the same woes of Web2.

U.S. regulators are also continuing to explore how genAI could impact various industries and consumers more broadly. Last week, the Federal Trade Commission hosted a virtual roundtable to address copyright concerns and other issues with various authors, artists and others. (Participants included Jen Jacobsen, executive director of the Artist Rights Alliance; Duncan Crabtree Ireland, national executive director and chief negotiator for SAG-AFTRA; and Umair Kazi, the director of policy and advocacy at the Authors Guild.)

“The ability to create synthetic voice with as little as three seconds of recorded audio is “easy, simple and dangerous,” said Tim Friedlander, president and founder at National Association of Voice Actors, who also participated in the roundtable.

“Synthetic audio can be used to deceive consumers into believing that a trusted voice is communicating with them,” Friedlander said. “At a minimum, consumers should know when AI is used to generate voices and receive a warning that the information they’re gonna receive may not be accurate. We’re also coming into an election and we know that synthetic voices will be used, they have been used. And it’s possible for anybody to use synthetic voices for disinformation. No one, a voice actor or anybody, wants to be the recognizable voice of disinformation in this coming election.”

Other news

  • The AI-assisted design race continues to gain more momentum. Last week, Canva debuted a new “Magic Studio” that includes new AI tools for generating text, videos, images and various design features. One of the most noteworthy updates is a new integration with Runway AI, a popular video generator, which gives Canva’s 150 million monthly users another way to create text-to-video clips and gives Runway a new route to scale beyond its existing users.
  • Adobe announced a new partnership with Google to let Chromebook Plus users have access to Adobe Express and Photoshop, which are both powered by Adobe Firefly. Adobe is expected to unveil more AI tools this week during its annual Adobe Max conference.
  • Social platforms are adding more AI features for marketers. On LinkedIn, a new tool called Accelerate will generate B2B campaigns within minutes including a draft of copy, images, campaign targeting parameters and ways to optimize. Meanwhile, Meta rolled out new ways for advertisers to generate image backgrounds, text for ads and adjust image sizes for each format.

AI hot takes

  • “The reason why we use AI that helps in my agency is I can now create renderings of my experiences. It allows me to go tackle and do better work and give clients the vision in terms of what I’m selling the same way [as] a Weiden [& Kennedy], an Anomaly, a Droga and all of that. So for me, the way I look at AI is leveling the playing field from technology and now I’m gonna go kick their ass.” — Coltrane Curtis, founder of Team Epiphany, said at last week’s Digiday Media Buying Summit.
  • “Copyright is not and cannot be the only tool to address the deep concerns creators hold about how their works are used. The Writer’s Guild of America has demonstrated the power of collective bargaining to secure important rights on how they will interact with, use and be subjected to generative AI.” — FTC Commission Rebecca Kelly Slaughter during the agency’s virtual roundtable
  • “[Generative AI] doesn’t change the process or the technology 100% overnight. But every month or two, we’ll find like another way that we can make something more efficient using AI. I was on the phone with our creative team today about a client brief. And they’re actually using AI to generate the first-round of concepts with a client that they’re going to show to a client. So instead of having an art director, write it down, they’re using AI to do that. It is changing things incrementally, but I think the increments are fast enough that it’ll be a big difference over the next couple of years.” — Kasha Cacy, Chief Media Officer of Known, who also spoke at Digiday’s Media Buying Summit

More in Media

Meta AI rolls out several enhancements across apps and websites with its newest Llama 3

Meta AI, which first debuted in September, also got a number of updates including ways to search for real-time information through integrations with Google and Bing.

Walmart rolls out a self-serve, supplier-driven insights connector

The retail giant paired its insights unit Luminate with Walmart Connect to help suppliers optimize for customer consumption, just in time for the holidays, explained the company’s CRO Seth Dallaire.

Research Briefing: BuzzFeed pivots business to AI media and tech as publishers increase use of AI

In this week’s Digiday+ Research Briefing, we examine BuzzFeed’s plans to pivot the business to an AI-driven tech and media company, how marketers’ use of X and ad spending has dropped dramatically, and how agency executives are fed up with Meta’s ad platform bugs and overcharges, as seen in recent data from Digiday+ Research.