To beat AI, you might need AI.
On Friday, DoubleVerify announced it has expanded its Meta media quality capabilities. DV customers can now evaluate Meta’s inventory for content categorized as “Inflammatory Politics and News” (IPN) within DoubleVerify’s system.
DV’s IPN category covers “controversial news and political topics,” according to the press release. You know, stuff that is “inflammatory,” of which there is no shortage of right now as we all contend with the fun that is the digital media ecosystem in an election year. (If you were still wondering, yes, your strange uncle’s post is definitely a conspiracy.)
Why This Matters:
Tucked away in DV’s announcement is the fact that the company’s IPN category now includes a new capability: evaluating and classifying deepfakes. Ultimately, artificial media has become much easier to create with Gen AI tools like ChatGPT and X’s Grok. The quality has also increased dramatically. See these examples. As a result of easy-to-use, better tools, deepfakes are a growing concern across all platforms.
This, of course, can be problematic, especially in an election year. In February, researchers said, for example, “Deepfakes… pose an alarming threat to society as they are capable of spreading misinformation and changing the truth.” Relatedly, a June study by Google DeepMind found that AI-generated political content is prone to misuse. AI giveth and AI taketh, basically.
The rise in malicious use of Gen AI and the proliferation of deepfakes on social platforms such as Facebook, Instagram, TikTok, and YouTube are growing concerns for advertisers. No brand wants their ads next to deepfaked or cheap faked content about Kamala Harris or Donald Trump, especially in a hypercharged environment. Understanding the trends and patterns around this content post-bid is important. Enter DV, which tries to give brands tools and data to navigate.
DV’s deepfake detection is powered by the company’s “AI classification engine,” Universal Content Intelligence. The tech relies on AI to analyze objects and people in a content’s visual elements, dialogue and music in audio and speech, and text—including metadata, captions, and transcriptions.
Experts React:
“As citizens around the world prepare to vote, giving advertisers the ability to choose the content they want to be aligned with is critical,” said DoubleVerify CEO Mark Zagorski. “By leveraging DV’s new AI-powered category, advertisers can protect their brand equity and evaluate suitability on Meta based on their specific marketing objectives.”
Our Take:
On the AI front, DV recently shared research on the negative impact Gen AI is having on app store reviews, helping to inflate them to encourage user downloads and attract advertisers. Gen AI tools have “numerous advantages,” according to DV, “yet it also poses significant challenges, particularly in the realm of advertising and marketing.”
The company is clearly focusing on AI, emphasizing its own capabilities and highlighting the challenges AI poses for advertisers, brands, and platforms. Those challenges are only growing and it’s kind of fascinating to see that the solution to combat them is also AI. In that sense, it’s a bit of an arms race. Good guys have more money but, often, the bad guys have more time.
