3 min read

Exploring The Power of Human-in-the-Loop AI

Exploring The Power of Human-in-the-Loop AI

When ChatGPT, by OpenAI, burst onto the scene in late 2022, the volume of coverage, discussions, and debates around AI and what it means for the future surged.

Unsurprisingly, during ShopTalk 2023 in Las Vegas, Nevada, numerous keynotes and session tracks focused heavily on AI and the applications, benefits, and (at times) shortcomings the spotlight tool may mean for individual brands and customer experience. 

In one session, Jeffrey Katzenberg, Co-Founder & Former CEO of DreamWorks Animation & Managing Partner, WndrCo, stated about the abundance of AI tools, “If we think digital transformation has moved quickly over the last ten years, wait to see what happens in this decade, between 2020-2030 technological advancements will have 10x the impact [compared to the last wave of digital transformation].” 

The general tone around AI throughout the show was cautious optimism, unearthing incredible applications for the evolving technology while still erring caution regarding the current state of generative AI. 

 

 

Is Generative AI like King Joffrey - Given the Iron Throne Too Soon?

British polymath Michael Polanyi notably stated, “We know more than we can tell.” And according to Ge Wang, an American computer scientist and Stanford Professor, this sentiment directly tracks what gives modern AI its allure, “Its ability to spot patterns in complex phenomena that defy rule-based descriptions means it can essentially understand them for us.”

While the concept of total automation can be appealing to many organizations hyper-focused on the outcomes of AI and the optimism around its bottom-line impact, today’s version of generative AI through tools like ChatGPT and MidJourney, still lacks the granularity of human control, which leads many experts to agree that total automation through generative AI will not be the endgame news frenzy is making it out to be.

For example, Katzenberg has been vocal in the past about the risks today’s generative AI could pose a risk for brands if accepted with blinders on:

 

Leaking sensitive data:

With generative AI, bad actors can mimic the writing styles of their victims by training on publicly available content. In turn, the doors to gaining access to accounts via phishing in the form of impersonation open.

 

Get into legal trouble:

Are you inadvertently using copyrighted material or intellectual property?

 

Offend your customers:

Fun, creative inspiration; great! Reliance on factual queries; is not such a good idea.

 

Large companies won’t adopt new technologies at a large scale if one of these brand-ruining scenarios is plausible.

While this may sound like a setup to side with the CEO of Twitter, Tesla, and SpaceX, Elon Musk, or other prominent names in the tech industry like Apple Co-Founder Steve Wozniak, who recently signed an open letter to pause giant AI experimentation, stating that advanced AI needs to be developed with care: “[In] recent months AI labs [have] locked in an out-of-control race to develop and deploy ever more powerful digital minds that no-one - not even their creators - can understand, predict, or reliably control."

That’s not quite it.

Instead, we’re asking the border question of not “How can we create more powerful digital minds,” but “How can we optimize the duality between automation and human interaction?”

 

Human-in-the-loop AI set to take over Generative AI Ambitions

“Generative AI, at least in its current state, is a creativity amplifier, maximizing the creative output of humans and allowing people to communicate their ideas more efficiently," says Evolv’s Chief Technology Officer, Tyler Foster. "I think you’ll see a huge explosion of great, human-generated ideas communicated better.”

The sentiment of human participation–commonly referred to as "Human in the Loop" (HITL)–is shared.  

Ge Wang, a professor at Stanford, says, “What if, instead of thinking of automation as removing human involvement from a task, we imagined it as the selective inclusion of human participation? The result would be a process that harnesses the efficiency of intelligent automation while remaining amenable to human feedback, all while retaining a greater sense of meaning.”

Creating a robust system that allows both sides to interact continuously precisely is what Evolv AI’s product roadmap focuses on and always has.

“A system like ours was really built for this era,” says Foster. “As we bring our next generation of generative AI to market, which will come in rapid succession over the next few months, it will all be human-in-the-loop.

Foster continues, “In general, as the pace of creating increases, traditional approaches to learning what’s working and what isn’t collapse quickly. Existing A/B tools, and some AI-led experimentation tools that are laggards in the industry, won’t be able to keep up.

While performing AI-led experimentation with Evolv AI, the technology works alongside Evolv’s fully managed team of AI experts and organizational stakeholders to set up correct hypotheses, analyze data in real-time and serve progressively better experiences based on a company’s approved knowledge, processes, and policies. This approach ensures brand safety and accuracy, meaning enterprises can enjoy the benefits of the AI revolution while avoiding some of its major pitfalls.

At ShopTalk 2023, Katzenberg underscored sentiments saying human-in-the-loop AI, or what he dubs “Sanctioned AI,” is the foundation for tools that will soon transform the retail tech stack, providing more reliable messaging around what your customers want. According to Katzenberg, “Sanctioned AI will be invaluable to retail - both in-store and online.”

Early adopters of human-in-the-loop advancements will have a head start. As Katzenberg says, this version of AI "Gives brands the ability to deliver not just customer experience but customer love.”

We agree.

Take your first steps into early adoption with Evolv AI’s intelligent digital experience optimization and personalization platform and enjoy not only the efficiency or correctness of AI technology but also human preference and agency. Learn more here.

Generative AI Trends and Use Cases to Follow

3 min read

Generative AI Trends and Use Cases to Follow

If you can, think way back to 2008 when Apple first launched its App Store. At its inception, there were a fraction of the downloads we have today as...

Design Thinking: Understanding the Impact of AI

4 min read

Design Thinking: Understanding the Impact of AI

When you think of design, you may think of something aesthetically pleasing. But good design is much more than that. It’s a functional experience...

Active Learning: The Intersection of AI and Design Thinking

3 min read

Active Learning: The Intersection of AI and Design Thinking

Imagine this scenario. The digital assets you’ve created as conversion points for your customers need improvement. You’re unsure what changes will...