Facebook Tests 3D Product Images in Marketplace Among Advanced AI and AR Shopping Initiatives

While Facebook takes a big leap into eCommerce with its new Shop options for both Facebook and Instagram, it’s also working on a range of more advanced AR and AI-based tools to improve the online shopping experience.
This week, The Social Network has outlined some of its evolving digital shopping tools, which are designed to make product recommendations more personalized, and address limitations in virtual product assessment.
First off, Facebook’s testing out a new 3D-like option which enables Marketplace sellers to provide a full view of their items within their listings.
As explained by Facebook:
“We’re introducing Rotating View, a state-of-the-art 3D-like photo capability that allows anyone with a camera on their phone to capture multi-dimensional panoramic views of their listings on Marketplace. This feature allows any seller with a camera phone to turn regular 2D video into a 3D-like interactive view. We’ve started testing this feature on Marketplace for iOS Sellers to start.”
One of the key challenges in not being able to physically view and hold an item is that you don’t know what you’re getting, and while questionable online shopping experiences have left many regular buyers helpfully skeptical, there are still challenges in ensuring that you’re getting what you’re agreeing to pay for. 3D visuals can help with this, showing any potential flaws or issues that you might not get with a selectively angled image.
Also on Marketplace, Facebook’s using a new, machine-learning based system to help sellers better tag their items.
“This new system, called GrokNet, automatically suggests attributes such as colors and materials when sellers upload a photo of an item for sale on Marketplace, which makes posting a listing much easier. And on the buyer’s side, predicted detailed descriptions provided by [the] system allow you to search Marketplace not just for black chairs, for example, but specifically for a black leather sectional sofa – even if the seller didn’t explicitly add those details to the description.”
GrokNet can also automatically provide product tag suggestions for Facebook Pages, which could help businesses using the new Shop option to maximize discovery of their items.
“When Page admins upload a photo, GrokNet can suggest potential products to tag by visually matching between items in the photo and the Page’s product catalog.”
That’s somewhat similar to Pinterest’s object identification tools, which can pinpoint specific items in an image to use for search and discovery across the platform – though Facebook’s system is more specifically aligned to the products uploaded by that Page.
That could make it easier to generate interest in your offerings – simply upload a photo with your products in it, as you normally would, and Facebook’s system will tag the products you have listed, and facilitate direct connection to those listings in your shop.
The process is currently in testing.
Looking even further ahead, Facebook is also testing new, AI-based tools that will be able to learn what style of clothing you prefer based on your existing wardrobe, then make product recommendations based on this.
“[We’re] prototyping an intelligent digital closet, which lets you take photos of your outfits and digitize each item within seconds. The digital closet can provide not only outfit suggestions based on planned activities or weather but also fashion inspiration based on products and styles that you like, so you can shop in the context of what you already own.”
That, also, sounds similar to Pinterest’s ‘Shop the Look’ Pins, which identify items to buy based on a photo – but again, Facebook’s system is a little different. In this instance, Facebook’s process is looking to utilize advanced matching systems to improve relevant discovery based on your existing wardrobe preferences, as opposed to what might look good on a model in an image.
Facebook also says that it’s working on developing new AR try-on type tools for products, which enable users to see how they look in, for example, branded sunglasses based on an ad or Page visit.
Instagram started testing these AR ads with selected partner brands late last year, and Facebook says that it’s looking to expand on this “to support more businesses and products in the future”.
The key to all of this, of course, is simplicity. Facebook shopping won’t work if it’s difficult for merchants to upload their product catalogs, on-platform buying won’t work if the experience isn’t great, and convenient, for consumers. Facebook knows this, which is why it’s taken so long to make the next leap into broader eCommerce offerings.
Enabling on-platform buying makes sense, but if it’s not easy, on all sides, it will clutter Facebook and Instagram, and potentially turn users away due to over-commercialization.
But if Facebook can get it right, it could be huge – and these new tools in testing point to the next phase of that shift, when eCommerce becomes a more integrated, intelligent way to shop.
Trump Hasn’t Been Arrested But But AI Images Are Fooling People

The former U.S. president Donald Trump still walks free as of the time of writing. But certain AI-generated photos on Twitter tell a different story. These deepfakes depict a world where one of the most powerful white men in America can be treated like any other citizen, meaning actually be held accountable for his actions. Except they’re not real. Some people believed they were, and that’s alarming for anyone who cares about media literacy.
This weekend, Trump told his supporters that he was expecting to be arrested on Tuesday over allegations of hush money paid to the former porn star Stormy Daniels. To be clear: The case against him exists. The Manhattan district attorney’s office has previously asked Trump to testify to a grand jury. Both legal minds and journalists are still speculating over whether or not he’ll be charged at all. So this whole saga is still far from a done deal. That hasn’t stopped Eliot Higgins, a citizen journalist, from using Midjourney to make these AI generated images of Trump being tackled and chased by the police.
While the original tweet has over three million views, these images have been shared across Twitter, TikTok, and Facebook, often without context or taking them at face value. Twitter hasn’t labeled these images as fake, so some people are having trouble with identifying them as machine generated (One tweet that pointed to this problem garnered over a thousand likes). Facebook is similarly chaotic. Some users chose to disclose the photos’ AI origins, and some didn’t. Many TikToks aren’t disclaiming that the images are AI-generated.
This is a huge problem when right-wing grifters like Ian Miles Cheong are using them to galvanize Trump supporters to action. New York City is already preparing for the former president’s supporters to riot in the streets over a possible arrest. These fake images of their “imperiled” leader just add more tinder to the fire. Kotaku reached out to ask Higgins for a comment, but did not receive one by the time of publication.
G/O Media may get a commission
Look, it doesn’t matter if you’re savvy enough to tell whether or not these images are real. Outside of our social media bubbles, large swaths of the country are vulnerable to misinformation and fake news. These images have already made their way to Facebook, a platform with millions of vulnerable users. As long as Americans want a Trump arrest to be real, such images will continue to go viral.
Hey, local news publishers: Give the people a calendar

Blairstown, Paterson, and Trenton are three very different communities in New Jersey, but when Sarah Stonbely, the research director of the Center for Cooperative Media at Montclair State University, surveyed residents about what they need from their local news outlets, she found they had a number of needs in common.
Paterson, one of New Jersey’s largest cities, is majority-Hispanic and also has a sizable proportion of Arab residents. (Paterson residents were surveyed in Spanish, Arabic, and Bengali as well as English.) It has a below-average median income for the state. Trenton, another large city and the New Jersey state capital, is roughly half Black. And Blairstown is a small, rural town that is more than 90% white.
But all three communities had lost most of their existing local news outlets over the years. All wanted more service journalism, in the form of information about municipal government meetings or contact information for local leaders. And all relied heavily on local Facebook groups for news, even though they also understood Facebook’s flaws.
Stonbely compiled her findings in this new report and shared them with hyperlocal news outlets that had recently launched in the communities: The Paterson Information Hub in Paterson, which is a news product of the nonprofit hub Paterson Alliance; the Trenton Journal in Trenton; and the Ridge View Echo in Blairstown. All three outlets are grantees of the New Jersey Civic Information Consortium, which we’ve covered here.
I asked Stonbely a few questions about her research.
Laura Hazard Owen: I am interested in your take on people’s impressions of the Facebook news in their communities. I feel like the way that we often hear about local Facebook groups is that they are tricking people, providing bad or biased coverage. But it sounds as if [the residents you talked to] know that these groups aren’t perfect and have mixed feelings about them.
Sarah Stonbely: I got that impression as well. I got the impression that people used [Facebook for local news] pretty grudgingly — they felt like it was kind of their best worst option, because of the drain of local news, plus people are already going there to see pictures of their friends’ kids and dog memes or whatever. I was very pleasantly surprised that they seem to recognize that it’s not ideal, it’s not necessarily “real journalism,” but they’re going to find out things there that they can’t find out anywhere else. They are sort of using it grudgingly because they don’t feel like they have a lot of other options.
Owen: I wanted to ask you about the logistics of doing research like this and getting people to actually show up. It sounds as if that was really hard: In Blairstown, for instance, “For the two scheduled in-person focus groups, 20 and 10 people, respectively, confirmed the day before that they would attend. Of those who confirmed, two people showed up for the morning focus group and zero showed up for the second, afternoon group.” I imagine it’s a common problem doing research like this, but do you have ideas about how you, or a different organization, could address that in the future?
Stonbely: Yeah, it was super frustrating, although not totally surprising. One thing that I think would be really helpful would just be to have more time built into a grant like this — to, for example, dig up more email lists. We were trying to get alumni lists from the high school, because many people who live [in Blairstown] have lived there for 20 or more years, but the high school wouldn’t give us those lists. I would just build in more time to figure out ways to reach people.
Owen: It seems as if a theme throughout was a desire for, like, calendars of municipal meetings — giving people more information about what is actually happening in their communities, things that they can attend. It seems sort of obvious. But you found that news organizations weren’t doing very much of that.
I thought that was really exciting. If you’re a publisher and you’re just in the weeds, starting a news organization and trying to do investigations or something, it just might not occur to you that [a municipal calendar] is something that would provide value.
The research was supported by funding from the Google News Initiative, and one condition of the grant was that “after the initial information needs assessments were complete, each outlet was to make improvements to their product based on the findings.” Here are the recommendations that Stonbely gave to The Paterson Hub, Trenton Journal, and Ridge View Echo.
Recommendations given to Paterson Hub
- The top two topics of interest for the community members we heard from were safety/crime and food (in)security, which do not readily lend themselves to events, which suggests that a different platform — perhaps an email newsletter or dedicated website — may be of more interest to those community members who want to hear about these topics.
- However, nearly half of people showed interest in events about exercise/recreation, housing affordability/homelessness, early childhood education, mental health, and music. This list of topics lends itself well to a shared calendar. In addition, the greatest share of survey respondents (more than half) said that they attend events that ‘help me solve everyday problems in my life’ and that ‘connect me to friends and neighbors.’ You can emphasize these types of events in your calendar — focusing on utility and connection.
- There is a long list of trusted organizations in Paterson; consider collaborating with these organizations on a calendar or news outlet (beyond just asking them to contribute content), so that trust is built in from the beginning. You may also tap people/offices on the list of most trusted sources.
- Engage to a greater extent on Facebook, and be present on Facebook groups that are relevant in Paterson; this is where most of the traffic is and where you’ll have the greatest visibility.
- Paterson is extremely diverse; take advantage of this diversity by offering your content in as many relevant languages as possible, but especially Spanish and Arabic.
Recommendations given to the Trenton Journal
- Create a dedicated section for posting the dates and times of upcoming municipal meetings, similar to your events page; publicize it via your newsletter and social media.
- Go one step further and cover municipal meetings regularly, even if it’s simply by providing a transcript.
- Consider adding a section on your website that lists all city and state departments, the services they provide, and their contact information.
- Consider offering different sub-pages for each ward, that can be tailored to the differing concerns and interest in each.
- Engage to a greater extent on Facebook, and be present on Facebook groups that are relevant in Trenton, especially Trenton Orbit and Peterson’s Breaking News of Trenton.
Recommendations given to Ridge View Echo
- Create a dedicated section for posting the dates and times of upcoming municipal meetings, similar to your events page; publicize it via your newsletter and social media.
- Go one step further and cover municipal meetings regularly, even if it’s simply by providing a transcript.
- Get more involved on Facebook, both on the feed and in groups.
- Continue to cover feel-good lifestyle issues in addition to hard issues.
- Consider adding a section that allows people to recommend service providers; maybe service providers could recommend themselves for a fee (similar to advertising but in a dedicated section)? Could list it as a Directory similar to the others that you have under Resources.
You can read the full report here.
Fans unearth Austin Reaves’ brutal old Facebook post about LeBron James

Storyboard
-
SEARCHENGINES6 days ago
Google Says Ignore Spammy Referral Traffic
-
PPC6 days ago
How to Get Found Online: Our Top 9 Tips for Local Service Businesses
-
MARKETING5 days ago
12 Best Practices to Boost Your TikTok Ad Performance
-
AMAZON1 day ago
The Top 10 Benefits of Amazon AWS Lightsail: Why It’s a Great Choice for Businesses
-
MARKETING6 days ago
A treatise on e-commerce data security and compliance
-
SEARCHENGINES6 days ago
John Mueller Offers Hreflang Google SEO Advice
-
MARKETING6 days ago
Top 10 Content Marketing Tools that Improve the Marketing Team’s Productivity
-
SEO7 days ago
SEO for Multiple Locations (Beginner’s Guide)