Box and Adobe Talk Next-Gen Work Beyond the AI Horizon

At a dinner last Thursday night in the SoHo neighborhood of Manhattan, Aaron Levie, CEO of Box, and Adobe’s Chief Strategy Officer, Scott Belsky, discussed with a group of reporters some real-world implementations of AI and how it may affect next generations of work.

“We’re in the start of a pretty massive sea change in technology,” Levie said, speaking on AI’s ripple effect that is top of mind across sectors. While there is plenty of excited speculation, and concern, about the further potential of the technology, grounded thinking about AI is bringing together working groups, task forces, and councils to evaluate models and compliance, he said. “There’s a lot of questions and there’s a lot of things the real world is now going to hit and face.”

Box has announced Box AI in private beta, which Levie said connects documents and content in Box securely to large language models (LLMs), whatever AI users want to work with. For its part, Adobe was thinking about generative AI as early as late 2019, Belsky said. Those initial thoughts led to an internal model that would evolve into the Firefly generative AI model for creating images. That and other resources Adobe has developed have helped creatives with potential timesaving automation for such tasks as choosing color palettes for illustrations, Belsky said.

“My forecast is that I think every piece of creative work will have generated pixels in it of some kind in the future,” he said. “I think that we’re about to enter a world where digital experiences are far more personalized.” Users may be able to make one million variants of one visual asset, Belsky said, then send them out for email or social media campaigns. However, this might also lead to some brands flooding the market with AI-generated content in unsustainable ways, he said, citing companies producing 1,000 blog posts per hour just to help the brand get noticed for SEO purposes.

“That’s not going to last,” Belsky said.

AI and Layoffs

A recurring concern about AI is the fear that the technology could lead to layoffs, especially among creatives. This year’s strikes by writers and actors in movies and television include disagreements with the studios over how AI might be used within the industry.

For instance, actors pushed back against one studio’s desire to capture the likenesses of background actors, own those images, and use them into perpetuity after only paying the actor just once. Writers worry studios want to rely more on scripts developed through generative AI with few human writers involved in the process. The studios have countered that the technology should be explored to remain competitive, but now face legal speedbumps about whether content created via generative AI is copyrightable.

Belsky said the way organizations might hire more staff for their creative and content teams, rather than reduce headcount, with the support of AI. “When you can get more ingenuity per person, you actually want more people,” he said.

Levie pointed out it is natural to see some camps resist, if not insulted by, the automation of work they learned to perform, while others might see it as a way to be more productive. “How do we reconcile these different views?”

Belsky cited that when digital photography emerged, some professional photography organizations would not let members join unless they worked with analog. “It took years for digital photographers to be admitted to shows,” he said. There will be changes in work to navigate, for example if automation does completely overtake specific tasks once performed by a person. “If your job was simply to turn a square banner into a rectangle banner all day, every day, and that can be done automatically now, you’re no longer going to be doing that job,” Belsky said. “The question is, ‘What will you be doing instead?’”

If that task was the extent of the individual’s current capabilities, they would be out of work, he said. If they do have other interests and skills, they might have a job creating variations on what automation produces. “Instead of just resizing them, how about localizing them for different markets? How about instead of A-B testing two, A-B testing 12, or 30, or 50? That’s what we’re seeing people do now that they have this technology,” Belsky said. “They’re just using their time on different things.”

AI technology could clear the way for such higher order tasks that can be performed, he said, which may elevate digital experiences. This is an idea others, such as Neil deGrasse Tyson, have put forth in defense of the opportunities that AI can bring by freeing up time on mundane, repetitive tasks to focus on more specialized work. “That’s a consistent theme of technology,” Belsky said. “The engineers have gotten far more efficient every year for decades, yet people keep hiring more engineers -- but they are hiring engineers to do different things. They’re not manually writing every little key character of code. There are repositories, there’s GitHub, now there’s GitHub Copilot. These technologies make us more effective and allow us to focus our human time on things that are going to require a uniquely human intellect.”

What to Read Next:

Generative AI Faces an Existential IP Reckoning of Its Own Making

AI-Generated Works Cannot Get Copyright: US Court

Risks and Strategies to Use Generative AI in Software Development