AI Briefing: Klarna hopes visual search in its e-commerce platform will help shoppers bridge in-person, digital gap

Visual search certainly isn’t new. In 2017, Google, Pinterest, Microsoft and Samsung all debuted AI-powered visual search tools for their platforms and devices. A year later, Snapchat released a way for users to scan real-life products and search for them on Amazon. (Speaking of Amazon, a 2019 report found the e-commerce giant had changed its proprietary algorithm to prioritize its own products.)

There’s still the challenge of how to spark behavioral change so people think to scan items when they want to find them. Although Klarna plans to wait before marketing its “Shopping Lens” with a big campaign, the company’s CMO David Sandstrom said he expects to see demand for it. For example, he mentioned Klarna’s recent survey of 3,000 shoppers, which found that 85% recalled wanting to buy something they saw someone else had, but 69% had trouble finding the item.

“It’s almost like a party trick product,” said Sandstrom. “It is very easy for influencers and creators to use the product in an educational but fun way.”

Klarna’s experimenting with AI in other ways, such as as using content creation and editing. It’s also built an “IP bot” to detect if an image resembles something that’s IP-protected. (The company already produces around 91,000 images per year and spends about 5,000 hours resizing and cropping images for various platform formats, Sandstrom said.)

The Shopping Lens is just one of several ways Klarna has built AI into its platform. In March, it debuted a new shopping assistant developed using OpenAI’s ChatGPT Plus, along with a recommendation engine to personalize products. Other AI tools are also in the works, he said.

Ultimately, Sandstrom thinks all these disparate elements will eventually converge. “This is only a wild guess,” he added. “What I see happening currently within the AI space [is] this is going to end up in everyone having a personal assistant in their pocket.”

In other AI news

• Misinformation from AI and other sources has quickly become a major concern for information coming out concerning the Israel-Hamas conflict. According to various reports, ChatGPT, Bing and Bard wrongly said there had been a ceasefire while others chats provided wrong information about electricity in Gaza. European Union officials are worried about AI-generated misinformation on social media including X and Meta.

• AI-generated videos and photos are also challenging newsrooms tasked with vetting information related to Israel and Gaza and the broader region. For example, CBS News CEO Wendy McMahon told Axios that AI-generated deepfakes and other misinformation are “flowing into our newsrooms at a scale, at a speed, and at a level of sophistication that will be staggering.”
  
• Adobe released a number of new genAI updates and debuted a new icon along with a redesigned “nutrition label” that will be embedded into the metadata of content created using Adobe software and other partner platforms. (Early adopters include Microsoft, Publicis Group, Leica and Nikon.)

• Democratic lawmakers are calling on President Biden to incorporate the White House’s “AI Bill Of Rights” when it issues an executive order on artificial intelligence.

• Google is the latest company to offer indemnity to companies that use its generative AI tools, following similar moves already made by Shutterstock, Microsoft, Adobe and various others. Meanwhile, another browser (Opera) has added new AI tools.

• The Harris Poll — a subsidiary of the Stagwell Marketing Cloud — released new AI-based research products called the Harris Quest. (The tools are part of the company’s efforts to integrate tech from Maru a year after the acquisition.)

• According to new findings from Digiday+ Research, ads on programmatic websites are “facing new brand safety risks amid the generative artificial intelligence boom.”

AI hot takes from humans

• “Economists who are critical of antitrust [would say]: ‘You’re all going after Google, you’re spending all this time and effort and resources going after Google. And as far as we know, tomorrow or in a year, OpenAI will make a deal with some advertising company and they’re gonna have the best advertising system ever. And Google’s gonna go out of business. Or nobody’s gonna use Google.’” — Eric Posner, University of Chicago professor and antitrust expert

• “What happens to your search input from the moment you input it is right now exclusively locked up 91% of time by Google. And they’ve prohibited Apple or anybody else who might have some AI features they’d want to bring into the search or the the browser bar…If you did break the exclusivity, if you did allow for innovation in search in the browser bar for Apple [or others], we would look back on this and say, ‘That was the moment when a new era of AI was brought into search. And maybe it’s 5 to 10 years later than it would’ve been had these exclusive deals not existed.” — Adam Epstein, president of adMarketplace

Advertising Week New York: An AI Preview

Nearly two dozen AI-related talks are slated during Advertising Week New York 2023, which kicks off today and runs through Thursday. Despite all the talk about how AI might improve creativity, the titles of most of the AI sessions weren’t very creative at all, sounding very much the same along the lines of “Unlocking AI” in some form or another. (One exception: Former Burger King CMO Fernando Machado’s AI talk titled “Think of a Color That Doesn’t Exist.”)

Here’s a glimpse at some of the AI-related talks over the next few days. (Extra fact: When Digiday asked Bard to analyze the AWNY schedule and identify AI-related talks, it accurately identified one, but then hallucinated and generated fake talks with two made-up titles and times.)

• Monday: 
  • Embracing AI: How Early Adopters Stay Ahead

• Tuesday:
  • How La Colombe and Kia Used AI to Improve Advertising Outcomes with LoopMe
  • AI for All: Democratizing Generative AI
  • Unlocking the Power of AI for Advertisers
  • Do More with Less: The Magic of Virtual Locations and Generative AI in Marketing
  • Unlocking the Potential of Generative AI in Advertising
  • Generative AI and the Changing Advertising Landscape
  • How Generative AI Powers Safer Contextual Marketing Choices to Reach Kids and Families
  • AI Ad-Ventures: Applying Machine Learning to Amplify Advertising Outcomes
  • Breaking Boundaries: AI’s Dynamic Role in Evolving Audience Engagement
  • Bots vs. Brains: A Stagwell AI Debate
  • Leveraging AI & Machine Learning to Elevate the Consumer Experience
  • Data, Trust and AI: What’s Powering the Future of Marketing

• Wednesday
  • Unlocking Advertising Success: AI-Powered Creative Performance Analysis
  • A Fashion-Forward Future: Using AI to Power Growth Marketing with Poshmark

• Thursday
  • Remember the Human: Defining Community and Online Spaces in the AI Revolution
  • AI — Friend or Foe?