A model for software giants embracing AI: how Adobe completes "AI transformation" step by step

With the rise of AI, all walks of life are ushering in great changes. The software giant Adobe has seized the loopholes in artificial intelligence and described an example of turning an elephant in the AI era.

Three months ago, when AIGC platform Midjourney created lifelike portrait photography, the market thought it was the end of Adobe.

Faced with the challenge of AI, however, the creative software leader turned decisively and quickly. Instead of dragging down Adobe’s business, AI has become a catalyst for its revenue growth.

How did Adobe transform from an “AI big loser” to an “AI big winner” and complete the “AI transformation” step by step?

Image source: Generated by Unbounded AI

Grasp the main loophole of AI - copyright security

1. A meeting to determine market positioning

Nine months ago, Scott Belsky, Adobe’s chief strategy officer, joined two dozen top marketing executives in a conference room in New York to discuss how artificial intelligence could transform advertising, enabling them to turn a string of words into a visual ad.

At first, Belsky thought they were on the same page about the enormous potential of AI. But marketers insist that their reluctance to use AI image models trained on copyrighted content could face a class-action lawsuit.

The conversation helped Belsky and Adobe position themselves in the race to deliver market-leading generative AI.

The $219 billion software giant has chosen to slowly and steadily integrate GAI tools into its core design products such as Photoshop, Illustrator and Premiere. They argue that a cautious approach is necessary and not risk using an AI model that could trigger a class action lawsuit.

In recent months, Adobe has picked up the pace without rushing. In March, Adobe released a beta version of Firefly, an AI tool that trains on hundreds of millions of images from the company’s vast image library, as well as some publicly licensed content and public domain images whose copyrights have expired. Last month, Adobe integrated Firefly’s functionality into a beta version of Photoshop.

**Because the model was trained on licensed and public domain photos, marketers told Belsky they would now consider using it for commercial content. Adobe announced two weeks ago that it would offer intellectual property indemnification to customers in the event that they are subject to copyright infringement lawsuits over images generated by Firefly. **Firefly is currently in beta, but Belsky says businesses will be able to use it commercially “soon.” Adobe is currently working with brands including Japanese marketing agencies, IBM, and Mattel to explore how they can use Firefly in their work.

Adobe wants to be the white knight in the AI race, in contrast to Stability AI’s diffusion model, OpenAI’s Dall-E 2, and Midjourney, which scoured the internet to train models on millions of images, ignoring efforts aimed at protecting artists and Copyright law for their works.

2. A critical step: Acquisition of the upstart Figma

Scott Belsky, Adobe’s chief strategy officer, was the former founder and CEO of the Behance design community, whose vision is to protect the rights of creatives. Following Adobe’s $150 million acquisition of Behance, Belsky started in March as Adobe’s chief strategy officer and executive vice president of design and emerging products.

Over the past few years, Belsky has left his footprints throughout Silicon Valley. He has been an angel investor in companies such as Pinterest, Eight Sleep, Warby Parker, Uber, and Airtable. Belsky relied on his startup connections. Acquired the largest acquisition in Adobe’s history - the acquisition of graphic design company Figma for $20 billion. Belsky has been friends with Figma co-founder Dylan Field for many years.

However, the deal now faces scrutiny from the U.S. Department of Justice, which is said to be preparing an antitrust lawsuit to block the deal. With its $20 billion Figma acquisition under threat,** the software giant is turning its attention to Firefly, its copyright-safe generative artificial intelligence model. **

Joining forces with members of Adobe’s creative cloud, research and digital media teams, Belsky wants Adobe to protect creatives, many of whom are seeking compensation for AI models trained on their work. There are currently several infringement lawsuits against Stability AI and Midjourney, and there is no legal consensus on how courts should determine what constitutes a fair use of AI.

3. Steady “transformation” of copyright protection

In fact, Adobe’s generative AI efforts began in 2019 when Alexandru Costin, Adobe’s vice president of AI at the time, showed how the AI could tilt the Mona Lisa’s mouth into a smile or change her posture.

At the time, AI technology was behind Costin’s vision. He said no one could see how this blurry 64-pixel image could turn into a fully editable 4,000-pixel image two years later. Belsky urged him to be patient, the technology isn’t good enough to be integrated into Adobe’s flagship product yet, but its time will come soon. However, when Adobe launched the Firefly tool in March, it was already behind its peers.

Still, Belsky said user interest in Firefly has surpassed even Adobe’s highest projections, with users creating 500 million images with its tool in the three months since Firefly’s launch.

**Importantly, while companies like OpenAI, Stability AI, Midjourney, and Runway may be first movers, Adobe believes it has a crucial advantage: After watching other companies draw lawsuits for blatant copyright infringement, the company It is now possible to snatch the dominance of the market by offering more trustworthy products. **

**4. Will countries stand by technology companies? **

AI infringement is becoming a top priority, says David Hoppe, a lawyer at Gamma Law Firm in San Francisco:

This situation is developing rapidly, and a series of copyright infringement lawsuits against Stability AI and Midjourney - from Getty Images and a class action lawsuit from artists - may cause huge reverberations in the AI sphere. Intellectual property lawyers have two concerns with generative AI models: first, whether any images created by the AI models can be copyrighted; and second, whether these models violate copyright law by putting together images from the Internet.

** So far, the artists have not won the argument, and if Stability AI and Midjourney win their lawsuit, its models will continue to use existing artwork to generate new images. So for Adobe, it will be difficult for Firefly’s model to gain a dominant position. **

Belsky is betting that other AI tools will eventually have to account for strict copyright restrictions, at which point many will follow the path of music streaming service Napster.

The original music streaming emerged as a cultural force at the turn of the millennium, a moment of free play, without legislation or real rules and structures. It collapsed under the onslaught of record industry lawsuits, opening the door for new companies like Spotify to arrive with record label licensing deals. Firefly is Adobe’s way of jumping to the end of Napster’s corporate story, releasing a legally defensible, “moral” version from the very beginning of the rise of generative AI.

In the future, Belsky said, Adobe could strike licensing deals with entertainment companies so that Firefly users can use well-known characters or brands in their productions, much like YouTube and Spotify do with music companies.

For now, Adobe’s efforts to respect copyright are at risk. Last year, the U.K. government announced it would allow AI developers to train on copyrighted content, a statement it reversed in March after the music industry pushed back, reconsidering U.K. AI regulations). Soon, U.S. law may be on the side of the big-model companies, too.

Morgan Stanley: IT service providers will be the beneficiaries of the AI wave

**For IT service providers embracing AI, Morgan Stanley believes that in the short term, IT service providers may be the beneficiaries of data and AI-driven spending. **

Morgan Stanley noted that advances in deep learning applications, such as generative AI, have unlocked a variety of new jobs that can be automated. While previous technological advances have augmented or automated routine, rule-oriented tasks, AI will impact more complex tasks that require cognition, learning, and decision-making. Automatic generation of natural language, code and digital content and task execution could drive significant efficiency gains in professional services, including the more than $1 trillion IT services industry.

Morgan Stanley sees IT service providers as a whole as near-term AI beneficiaries, and believes that early adopters of AI tools internally can benefit from productivity and marginal benefits. Morgan Stanley expects this scope to be limited to early adopters, user productivity improvements, and IT service providers investing strategically to build out AI capabilities.