Adobe says its new generative AI will help creative interests, not hurt them.


In 2022 OpenAI’s Dall-E The service has surprised the world with its ability to convert text queries into images. Now Adobe has built its own version of the company’s widely used creative tools with tools that start with technological improvements.

On Tuesday, Adobe released the first two members of the new Firefly A suite of generative AI tools for beta testing. The first tool generates an image based on a text query, such as “A powerful alligator jumping out of the water during a lightning storm,” with hundreds of possible conversion results. Another applies speed-based methods to text, allowing people to create letters that are hairy, skinny, mossy, or whatever they want.

Firefly is currently available on Adobe’s website, but the company also builds Generative AI directly into other tools, from its Photoshop image processing software, Illustrator for Designs and Adobe Express to create instant videos. The company did not disclose the pricing approach for the new devices.

Creative professionals can look beyond mechanical tools like choosing colors and cutting videos to the heart and soul of their work as Firefox enters the creative realm. It’s not unreasonable for professionals to feel squeamish when AI shows new smarts like translating documents, interpreting tax codes, composing music and creating travel itineraries.

Like other AI enthusiasts, however, Adobe sees artificial intelligence as the latest digital tool to amplify what humans can do. For example, Firefly could eventually allow people to use Adobe tools to customize designs for individuals, rather than just creating one design for a large audience, said Alexandre Kostin, vice president of Adobe Generative AI operations.

“We don’t think AI will replace inventors. We think that inventors who use AI will be more competitive than inventors who don’t use AI. That’s why we want to bring AI to the fingertips of all our users,” Kostin said. “The only way to succeed in AI is to embrace it.”

Adobe Firefox products are compatible with the company’s library of stock images, public domain and licensed works. The company has worked to reduce biases in training data that AI models may reflect, such as the fact that business executives are male.

AI is a “sea change”.

Artificial intelligence uses processes inspired by the human mind to perform computer tasks. If-then programming is trained to recognize patterns in complex real-world data rather than following traditional and rigid rules. With advances in AI hardware, software, algorithms, and training data, the field is advancing rapidly and touching all corners of technology.

The latest flavor of technology, generative AI, can create new things on its own. The most famous example, ChatGPT, software can record, record conversations and write poetry. Microsoft is using the ChatGPT Technology Foundation GPT-4 to help boost Bing search results, provide email writing tips, and build presentations.

AI tools are popping up everywhere. Adobe has used AI for years under the Sensei brand for features like recognizing human subjects in Lightroom photos. Transcription of speech to text in Premiere Pro videos. EbSynth applies photo style to video., HueMint creates a color palette And LeiaPix converts 2D photos into 3D scenes.

But it’s the new generative AI that brings new creative possibilities to digital art and design.

“It’s a sea change,” he said Forrester analyst David Trugh.

Example Adobe Uses Generative Ai To Shape The Letter N With Mossy, Gold, Or Textured Or Thousands Of Red Particles.

One of the first in Adobe’s Firefly family of generative AI tools, it designs text based on queries such as “The letter N made of gold with intricate ornaments.”


Alpaca offers Photoshop plug-in to generate artAnd you can Aug X Labs Convert a text question to a video. Google MusicLM converts text to musicAlthough not open to the public. Dal-E has captured the internet’s attention for its often-stunning imagery — the name marries Pixar’s Wall-E robot with the real-life artist Salvador Dali.

Related tools like Medium trip and stable AI’s Stable distribution Expand the technology further.

If Adobe doesn’t offer generative AI capabilities, creative professionals and artists will find them elsewhere.

Of course, Microsoft Tuesday It integrates Dall-E technology with the Bing image generator. Service.

Training AIs isn’t easy, but it’s getting harder, at least for those with a healthy budget. Chip designer Nvidia announced on Tuesday that Adobe is using the new H100 Hopper GPU Firefly models by A A new service called Picasso. Other Picasso clients include photo licensing companies Getty Images and Shutterstock.

Legal engineering

Developing good AI is not just a technical matter. Adobe designed Firefox to sidestep the legal and social problems AI brings.

For example, three Artists sued Stability AI and Midjourney In January, their work on the use of AI training data. “They will strive to stop this blatant and massive infringement before their entire careers are destroyed by an overworked computer program,” their lawsuit said.

Getty Images also sued Stability AI.“He illegally copied and processed millions of copyrighted images,” he alleged. It licenses its huge catalog of photos and other images for AI training, but Calm AI doesn’t license the images. Stable AI, DeviantArt and Midjourney did not respond to requests for comment.

Adobe wants to make sure artists don’t have to worry about such problems. No copyright issues, no brand logos and no Mickey Mouse characters. “You don’t want to accidentally infringe on someone else’s copyright,” Costin said.

The approach is clever, says Trug.

“What Adobe is doing with Firefly is very similar to what Apple did 20 years ago with the introduction of the iTunes Music Store,” he said. At the time, Napster music sharing showed interest in online music, but lawsuits from the recording industry killed the idea. “Apple has jumped in and designed a service that allows people to access music online, but legally, easily and compensating content creators rather than stealing.”

Solve other AI problems

Adobe has also worked to combat another problem that can disrupt businesses, showing distorted or distorted images.

And it’s working to solve some other pressing AI issues. For one thing, Firefox is tagging images with metadata, which is the metadata generated by the AI ​​system. Content validity initiative The company helped. CAI’s original purpose was to write metadata to image files to help news publishers and others verify that photos are genuine and not fake or edited.

Among other things, Adobe is allowing creators to include a “do not train” tag on their images using CAI technology, or a tag instructing them to exclude other AI training systems from training data.

Compliance with the system is voluntary, as is the “robots.txt” file that website operators can use to exclude their sites from search engine indexing.

“Some may decide not to respect it, and there’s nothing we can do about that,” Costin said. But other AI training companies have indicated that they may adopt a system that honors the needs of some enthusiastic artists.

Now it’s up to Adobe to convince creatives that it’s time to catch the AI ​​wave.

“The introduction of digital innovation has increased the number of creative jobs, not decreased them, even though it seemed like a big threat at the time,” Costin said. “We think the same thing will happen with generative AI.”

Editors’ note: CNET is using an AI engine to generate some personal finance explanations that have been edited and fact-checked by our editors. See for more This post.

We offer you some site tools and assistance to get the best result in daily life by taking advantage of simple experiences