Meta, Google and Snap are preparing for generative AI, as the technology has caught on with consumers through products like ChatGPT, and they see it as a powerful tool in advertising and creator economies.

Last week, Snap CEO Evan Spiegel discussed the potential for generative AI to power the app’s augmented reality features. Meanwhile, Meta CEO Mark Zuckerberg talked about generative AI in similar terms, to generate “images and videos and avatars and 3D assets, and all kinds of different stuff.” And Google CEO Sundar Pichai said that the company was “extremely well positioned as AI reaches an inflection point.”

The CEOs touched on generative AI in their fourth-quarter earnings calls. The tech brands all reported dips in advertising demand, which affected their revenue outlooks. Google’s Search and YouTube businesses reported year-over-year ad sales declines of 2% and 8%, respectively. Meta’s ad business was down 4% year over year, and Snap’s revenue was flat year over year.

The tech companies are dealing with a downturn in digital ad markets. They are grappling with changes to how data is shared across apps and devices under new privacy restrictions online. AI and machine learning are among the tools that could fix ad targeting. AI is useful in developing models that plan marketing campaigns and measure ad performance. Generative AI, though, is a creative catalyst—a technology that conjures text, images, videos and digital renderings in seconds. It’s also a nascent and unknown technology that could be used for deepfakes and lead to unforeseen consequences in culture.

Related: A marketing guide to generative AI

Generative AI has caught on among consumers, too. Last year, OpenAI shot to popularity with DALL-E and ChatGPT. DALL-E translates text prompts into images, and ChatGPT writes fast responses to complex queries, like a more dynamic search engine. Google and Meta have already been working for years on similar technology, but they have not quite entered the consumer conversation. Google, for one, sees a threat to its search dominance after Microsoft invested $10 billion into OpenAI and started to incorporate the AI into its consumer products, like Teams and Bing.

Here is how Meta, Google and Snap are answering:

Meta’s AI advertising plans

Meta is “working on things that could really empower creators to be way more productive and creative across the apps and run a lot of different accounts,” Zuckerberg said. Now, Meta is looking at generative AI as a long-term goal, like its metaverse ambitions. But the technology is already seeping into ad products, as marketers rely on machines to slightly adjust ad creative and copy.

“We’re using it in creative, we’re using it in targeting,” said Marne Levine, Meta’s chief business officer, in a recent phone interview. “We’re making it easier for advertisers to test.”

Levine was talking about AI that already is taking over more of Meta’s ad decisions. AI also is powering Meta’s recommendations for content to serve users on Facebook and Instagram, including what Reels videos to serve. Meta has ad products like Advantage+, which are AI-enabled campaigns that look for the best audiences.

Meta is touting AI in ads as the key to evolving in the new privacy landscape, where it has less data to track consumers, so it needs more sophisticated AI models to fill in the gaps. Last quarter, Meta said advertisers saw 20% more conversions on ads, with outcomes like clicks and sales, compared to the year before.

Generative AI could be another creative tool that has applications in ad products. For instance, it’s easy to envision marketers using text prompts to tell Meta’s ad manager to design an ad. And messaging is one of Meta’s fastest-growing ad products on WhatsApp, which topped 2 billion monthly users last quarter. Businesses are using messaging ads to send products and offers to consumers, and Meta said “click to message” ads are now at a $10 billion annual revenue run rate. It’s possible that AI-generated text services could help businesses chat, too.

For now, generative AI is being used in small ways to tweak ads, Levine said. “What we’ve done is we’ve leveraged AI investments to launch a new feature that gives advertisers the ability to apply small enhancements, like adjusting brightness, aspect ratio, to test placements, to improve performance of creative,” Levine said. “And so, what this means, is that advertisers can get smarter, faster, on what campaign creative combinations drive the best results.”

Meta has more advanced projects in the works for consumers to test AI-generated content. Last year, Meta teased a service called Make-A-Video, which has limited availability to the public. The service could spin up any type of artwork a person requests, say, a dog flying in a cape, or a spaceship on Mars.

Related: AI tools raise safety concerns for most consumers

Google’s all-in on AI

Last week, Pichai opened Google’s earnings call with a litany of AI developments. Google especially feels the pressure in this category, because the company has been investing in AI for decades, but its projects have been more walled-off from the public. AI already is used in Google products and ads, and Google is becoming more vocal about it. In fact, Google will note in its Super Bowl commercial for the Pixel 7 that its phone’s camera relies on AI for its advanced editing features.

Google has been positioned as an “AI-first” company for six years, Pichai said, and its research has been the “basis of many of generative AI applications you’re starting to see today.”

Google, like Meta, has infused its ad platform with AI training to help marketers pick the right creative, set campaign goals and target consumers across the web. “For our advertising partners, AI is powering dramatic campaign improvements and value-adding features for them,” Pichai said.

AI also is affecting Google’s core search product. Microsoft has been rushing to get ChatGPT into Bing, and Google has now accelerated its plans. It’s possible Google was being more cautious about unleashing a nascent AI brain for consumers to poke and prod. Now, Google is getting ready to uncage LaMDA, a language model that is similar to ChatGPT.

“These models are particularly amazing for composing, constructing and summarizing,” Pichai said. “They will become even more useful for people as they provide up-to-date, more factual, information.”

Snap imagines AI-generated pirate ships

Not to be outdone, Spiegel also sees the creative possibilities of generative AI, especially around Snap’s augmented reality projects. Snap’s Lens Studio has 300,000 creators building augmented reality filters, which are animated 3D digital creations. AR is still a nascent technology, mostly used to play games, to dress-up selfies with virtual decorations, and for more interactive shopping. Snap, Meta and other tech companies, though, are developing mixed reality devices that could make AR features available anywhere, through glasses.

“A lot of our most sophisticated AR lenses use generative AI technology,” Spiegel said. “And we also see a lot of opportunities just to make our camera more powerful with generative AI. I mean some simple examples are like improving the resolution and clarity of a Snap after you capture it or even much more extreme transformations or editing images or Snaps based on a text input.”

Generative AI is “critical” to the development of Snap’s AR program, Spiegel said. “There's a limited number of 3D models that have been created by artists,” Spiegel said. “And we can use generative AI to help build more of these 3D models very quickly, which can really unlock the full potential of AR and help people make their imagination real in the world.”

Eventually, generative AI could give anyone the power to create in augmented reality. “Imagine playing around with your kids wearing AR glasses and pointing, ‘Oh my gosh, there's a pirate ship and a big monster,’ and we can bring those to life using generative AI,” Spiegel said.