Google, Meta, and Amazon’s next frontier: AI-generated ads

Illustration: The Verge

Google is planning to use artificial intelligence to help companies create ads, according to a report from CNBC. Internal documents viewed by the outlet suggest that Google wants to use its new PaLM 2 AI language model to help advertisers generate assets that they can use in their ads.

This tracks with a report from the Financial Times in April, which similarly stated that Google could soon create ads by putting together images, video, and text supplied by the advertiser. That’s not the only way Google could use its AI language model, however. CNBC reports that Google is also looking into ways to leverage its AI model to provide video ideas to YouTubers. It could also integrate an AI chatbot within the Play Store, Gmail, and Maps to provide customer support.

Other tech giants, like the Facebook owner Meta, are looking to make AI toolsets available to advertisers as well. Last week, Meta launched the AI Sandbox, which acts as the company’s “testing playground” for the early versions of AI-powered advertising tools. The company is currently working on tools for background generation, image outcropping, and text generation that helps advertisers create “different messages for certain audiences.”

Meanwhile, a report from The Information indicates that Amazon is working on a way to generate photos and videos for companies looking to create advertising campaigns on the platform. The company is also rumored to be building a conversational AI search tool that’s supposed to answer questions from customers, help them compare products, and more.

Google held its annual I/O event last week, where it took the wraps off of PaLM 2 and a flood of other AI features powered by the model. In addition to rolling out AI-generated summaries in Search, Google is also working on an AI notebook that’s trained on your documents as well as a way for Play Store to build out their listings with generative AI.