AI deepfakes are a train wreck and a sellout for Samsung

On Thursday morning, I attended a Q&A panel with four of Samsung’s top smartphone executives. Until 2025, Samsung was the world’s largest smartphone manufacturer, and thus the world’s largest camera maker. It is still the second largest company after Apple.

Samsung handed me the microphone first. I asked:

We’re seeing a divide in society between people who want AI to do impressive things with their photos and videos, and those who don’t want AI to do that. anything With photos and videos because Eroding our capacity To believe that what we saw is real, Destroying the concept of photographic evidence.

Metadata tools like C2PA have Completely failed to stem the tide. Does Samsung have any new and different ideas on how to prevent AI images from taking over the world?

The four Samsung executives didn’t have any new and different ideas to share.

I’ll give credit to Won Jun Choi, COO and head of R&D at the mobile phone division, for not dodging the question. He told the chamber that the erosion of reality is a problem and he wants to fix it.

But he and other Samsung executives noted that the company needed to balance the desire for photorealistic photography with allowing smartphone buyers to be “more creative.” They passed the buck by pointing out that it’s an industry-wide problem, one that requires a broader conversation, and suggested that Samsung has already partially solved it simply by adding a watermark to AI-generated photos. Watermark can be removed easily.

Later, one executive suggested that our feelings toward AI-generated content may become more positive in the future.

Here are some of those answers.

“We recognize the problem,” Choi began, “because a lot of the content is generated by artificial intelligence.” “On the one hand, people want to be more creative, so we think we have to provide a solution so that people can become more creative,” he said. “On the other hand, nowadays it’s very difficult to differentiate between real and fake photos and videos. I think that’s a problem, and we’re aware of that, and I think it’s a problem that we have to solve at the industry level.”

“You might look at C2PA as a failure, but it’s still enough to provide a mechanism if people want to verify that those photos and videos were taken by AI. I think we have to provide a mechanism so people can use it. I think it’s an ongoing effort across the industry to solve this problem,” he added.

He concluded: “I think that if we make joint efforts to solve this problem, I think we should be able to do it.”

As does my colleague Jess Weatherbed he wrote earlier this weekwe’re starting to worry that this kind of “industry will solve this problem together” rhetoric, and C2PA itself, are a substitute for actual meaningful action.

And maybe respond to Our Samsung AI watch reportDave Das, CEO of Samsung America, also said that the company is still learning about the acceptability of artificial intelligence when creating its own ads. The company “just started using some AI content in our creatives, and the feedback has been very clear,” he admitted.

“We’re trying to figure out where to use it, and how to be very clear about when we’re using AI-generated content versus naturally generated content,” Das said.

But Das still sees it as a juggle between business priorities, not a social responsibility. “It boils down to giving the creator the choice,” he said, noting that Samsung’s mission is to find “the right balance.”

Later in the morning, a KTLA-TV technology reporter Rich DiMuro I sensed Samsung’s position with a completely different question. He wondered if Samsung might want to do this Easier For customers to remove their AI watermark from AI generated images.

“If I’m creating a Christmas card using all your fun tools, do I really want to say ‘AI-generated content?'” DiMuro asked.

Drew Blackard, senior vice president of mobile product management at Samsung America, led the charge. “If we feel that consumers really want that ability to remove watermarks, and there are other ways to address the first point about authenticity at the same time, then we will do both and deliver the best of what people want from experiences,” he concluded.

“At least for now, there’s enough consumer concern about authenticity, which is the main thing we’ve solved, in terms of watermarks in both the metadata and on the image itself. And not all services do that,” he said.

Blackard also suggested that we might look back one day and realize that AI-generated content is no big deal, and that our perceptions may change over time, in the same way perceptions of user-generated content changed in the beginning. He said people were concerned then about “the amount of content now flowing into the system” and how it wasn’t the professional video product they were accustomed to.

I wonder if Samsung and its fellow smartphone makers have thought the opposite: that AI-generated image perceptions might become equal. less This technology becomes favorable after people lose their jobs to cheaply produced artificial intelligence, and after lying, cheating, and stealing become much easier in the absence of reliable recorded evidence.

I wonder if they have thought about solving this problem “industry wide” before They break the dam, and whether perception might overturn against Samsung to help create the flood that comes next.