Google has spent the past few weeks promoting generative AI tools that can summarize search results for users, help them draft essays, and swap out overcast skies for sunshine in otherwise perfect family photos. Today it’s showing off what similar tools could do for its core business—selling ads.
New generative AI systems for advertising clients will compose text on the fly to play off what a person is searching for, and they'll whip up product images to save them time and money on design work. The features add to the swelling ranks of AI-based text and image generators that have been introduced to online services over the past few months, since the abilities of ChatGPT and its image counterpart DALL-E inspired global excitement about generative AI.
As the world’s top seller of online ads by revenue, Google has been using AI programs for years to help clients target users, as well as helping them design ads, like by automatically editing the size of images. Now, with more powerful AI models capable of tasks like generating photo-realistic images, it hopes to show that its ad business, which accounts for 80 percent of its total sales, can be more compelling to advertisers too.
The recent onslaught of AI-related announcements by Google has rallied shares of its parent company, Alphabet, suggesting that fears have diminished about the advent of ChatGPT-style web search crippling Google’s search and ad businesses.
Google is offering the new features to advertisers for free, but they could increase its revenue if AI-generated text and images encourage businesses to place more ads, or can draw more clicks from consumers. Google’s dominant role in online ad sales means the industry could be one of the first to broadly incorporate generative AI into their workflows. “We're able to deliver more relevant, beautiful ads to users, offer more creative freedom for advertisers, and deliver better performance,” says Jerry Dischler, the vice president overseeing Google Ads. He declined to discuss specific financial prospects for generative AI in ads.
As anyone who has experimented with an AI chatbot or image generator knows, their output can be unpredictable and even distasteful. And they have raised public concern over whether their development benefited from copyright infringement.
Dischler says the company will be “diligent” in monitoring the quality of images and text generated by the new features, some of which are available to advertisers in beta form already. Google is launching some of them more broadly than its top rival, Meta, which announced earlier this month that it was initially inviting select advertisers to try out its own generative AI features.
Offering generative AI in ads is likely expensive, because the computing costs to operate text- and image-generating models is sky high. At a conference last week, Meta AI executive Aparna Ramani said generating an output from those kinds of models is 1,000 times more expensive than using AI to recommend content and curate users’ News Feeds.
One of Google’s new features out now adapts the text of English-language search ads based on what a person typed into the company’s search box and Google’s data on the advertiser. Previously, each time a person searched, algorithms would have to select text to display from a collection an advertiser had manually written in advance.
With the text generation option, a search for “skin care for dry sensitive skin” could trigger an ad for skin cream with the auto-generated text “Soothe your dry, sensitive skin,” Dischler says. That may not seem revolutionary, but making ads more closely match searches could increase the chances of users clicking.
Google is also using its text-generation technology to offer a chatbot that ad customers can use to get suggestions for search keywords worth advertising against and text to go in those ads. “We would love to be able to offer personal support to millions of advertisers, large and small, but we think this is the next best thing,” Dischler says.
The Help Me Create My Ad prompt draws on Google’s data on past campaigns and analysis of a customer’s website and will be available to select US advertisers in English in July. Its design is similar to the company’s Help Me Write feature being tested in Gmail and Google Docs, which WIRED’s review found to be a good creative aid but also sometimes a stuffy writer that perpetuates stereotypes.
While the new ad-text capabilities mostly improve upon existing features, a new generative image tool offers ad buyers a more substantial upgrade to Google’s offerings. It crafts images that can be inserted into what Google calls Performance Max ads, which appear on Google apps and websites selected by Google’s algorithms. A screengrab of the tool provided by Google shows four photo-real images created in response to the prompt, "Ingredients used for pet food, like fish, chicken, raw meat and vegetables on a stylish slab-style plate, in studio lighting.” Testing in English will begin later this year, Dischler says.
Image generators can create new images by remixing patterns learned from huge collections of photos scraped from the web and other sources. Although tech companies have generally argued that using images that way does not breach copyright, recent lawsuits have questioned the practice.
Dischler insists the new feature should not burden advertisers with any legal liability but he describes navigating licensing in generative AI as an ongoing conversation. “Everything from our end has authorized use,” he says. Google allows intellectual property owners who suspect unauthorized use to file claims, and it takes down ads it finds in violation.
Image generators are also known to reproduce biases in their training data, with one result being that systems trained on web images that show more people of lighter skin tone generate images with the same tendency. A Google screenshot shows four generated images of cats—all orange or white—purportedly created by remixing the advertiser’s own content.
Dischler says that Google has designed its system to deliberately emphasize representation and inclusivity so that generated images reflect demographic diversity. “Ultimately, it’s up to the advertiser to approve new creative assets. It has to be policy compliant, and it has to be relevant to the user,” he says.
Google is restricting what advertisers can ask of the image generator, in line with the company’s AI principles, to avoid unjustly harming anyone, so they can’t run loose with it and try to unleash totally inauthentic or inappropriate ads. Dischler says the tool aims to allow smaller advertisers to produce more relevant ads at lower cost, and bigger advertisers to get bolder and more detailed images.
A second image-based AI tool will allow companies advertising or listing products that appear in search results or other Google services to use typed prompts to generate new background scenes for product images, remove existing backgrounds, and to increase their resolution. Called Product Studio, the tool will launch in the US later this year.
Google is also adding product ads to the new AI-generated search results that it began testing this month. An example provided by the company shows a list of AI-generated recommendations for “hiking backpacks for kids,” with the first option tagged with a “sponsored” label.
How well those ads sell could hinge on whether search users trust AI-written results, such as the shopping recommendations, and click through to spend money. That could require Google’s AI to improve its selections. In a demo at Google’s I/O conference earlier this month, its AI search recommended an electric bike that WIRED reviewer Adrienne So rated poorly. She says the kids’ backpacks that Google’s algorithms suggested are also odd choices when better options exist. Google spokesperson Lara Levin says suggestions are based on the world's most comprehensive data on products, including reviews of them, and even traditional search results may not align with a particular review.
Consumers are likely to see a growing number of AI-generated ads over coming months—whether they realize it or not.
Meta’s competing generative AI tools for ads announced this month can suggest variations of ad text, generate backgrounds for product images based on typed prompts, and auto-adjust the sizing of ads to fit on different apps.
Smaller players in digital advertising will almost certainly follow suit with similar offerings. Amazon is hiring for a team to work on advertiser-focused generative AI projects, according to job postings first reported by news outlet the Information. Like Google, Microsoft and Snap have both invested heavily in generative AI tools for consumers, such as chatbots, and they may see providing the technology to their ad customers as a natural next step.
