Creativity and humanity in the AI age

In the world of AI, 12 months is an eternity – so much was clear from the exciting new tech on display at the recent 2024 Rakuten Optimism conference.

One panel discussion titled AI Accelerates Innovation: The New Frontier of Creativity kicked off with a video introduction from two hyper-real AI avatars, who gave the Optimism audience an effortless welcome in 10 different languages.

The avatars then relinquished the stage to their flesh-and-blood counterparts – Rakuten Group Executive Vice President and Chief Marketing Officer Naho Kono and Samurai Inc. Creative Director Kashiwa Sato, the creative director and graphic designer behind many iconic Japanese brands such as Seven-Eleven, Uniqlo and Rakuten.

The session served as a follow-up to a discussion from Rakuten Optimism 2023, when Kono and Sato were joined by Kyoto University physicist Professor Koji Hashimoto to discuss the potential of AI tools. This year, Kono and Sato welcomed Professor Jun Suzuki, director of the Center for Language AI Research at Tohoku University.

From potential to reality

“It’s been barely two years since generative AI really started taking off,” Suzuki told the panel. “We can do so much more now. And I think you can all imagine that it will continue to improve drastically.”

The previous year’s discussion saw the panelists marvel at the text generation capabilities of AI, revealing that they had used a large language model to come up with the title of the conference session.

This year, Kono took the opportunity to show off a range of areas in which Rakuten has already begun effectively applying AI tools for business. These included advertising campaigns for Rakuten Beauty and Rakuten Fashion Week, an AI character to promote Rakuten Card, and a simplified ID verification process for Rakuten Mobile.

“This time last year, we were talking about potential applications. Now we’re actually here, applying AI,” Sato commented. “In this case, it’s not that the output has changed drastically, but rather the tools. So it doesn’t look like anything has changed from the customer’s perspective. But going forward, as the tools change, so will the output and ideas. I’m looking forward to that future.”

Suzuki researches artificial intelligence, natural language processing, machine learning and how computers are able to gain understanding of human language. He expressed a belief that in the near future AI will impress us even further.

“In the future, AI could become more interactive. We could make things in a dynamic way,” he predicted. “When you generate things in advance like we’re doing now, you have the chance to fix them up. But if you have to make it on the fly, you’ve just got that one chance. So we’re not quite there yet. In the future, I think we will be able to do this.”

Can we trust AI?

“What surprises me the most is that we know if we input something into an AI, we get a certain output,” Sato remarked. “But when it comes to the operating principles of how it gets to that answer, we don’t yet have a good understanding.”

“Even as we’re able to do dramatically more with AI, it’s becoming clear that there are a lot of potential problems,” Suzuki agreed. “One of those problems is that the generative AI we use now is often described as a black box. Even if we know that a certain action causes a certain result, we still don’t know very well what kind of logic it’s following. And as long as we don’t understand that, we won’t really be able to take full advantage of AI.”

Much of Suzuki’s research focuses on how well AI aligns with human sensibilities, looking at metrics such as whether AI chooses the same words to describe certain things.

“In actual fact, AI isn’t processing these things in the same way as a human. Of course, what’s happening in a computer is different from what’s happening in a human brain, so there’s a difference there, even if we’re seeing similar things come out.”

Suzuki says that even the biggest models aren’t quite there yet, but the path forward is clear.

“When there is a difference between how humans and AI feel about something, we aren’t able to fully trust our use of AI,” he added. “Only by making AI that aligns as much as possible with human sensibilities will the world be able to use it to its full potential.”

Staying optimistic in the AI age

In a world flooded with AI-generated content, what will become of human creators?

“Young creators and students often ask me what they should do going forward,” Sato told the panel. “It’s not a question of skill, but of morals, of philosophy – what does it mean to be human? It’s important that we face this question head-on.”

The flipside of this, Suzuki highlighted, is that AI tools can empower regular people to become creative.

“People who weren’t necessarily creators are now able to ‘make’ things,” he noted. “Now that the playing field has been leveled somewhat, you should be thinking about how to infuse your own personality into your creations to add a unique flavor.”

Sato concluded the session on an optimistic note.

“I’m mainly excited about AI. I want to see new things. I want to make new things. How we make things is going to keep changing – we could see creations that humankind has never seen before. And that’s something I want to witness,” Sato concluded. “If we do this session again, I’ll be excited to see what awaits us in the future.”