Facebook’s New AI Tool
For anyone who has ever dreamt of being able to click on any image and buy it. “This is something I’ve wanted to build since I watched the movie Clueless” Tamara Berg, a Facebook research scientist, told The Verge. “They really imagined everything back in 1995, but the technology I think is now finally ready to make it come to life.” And as the Clueless reference suggests, these ideas of shopping aren’t exactly new ideas. Even outside of Hollywood movies, the world has been trying to create and test technologies like this for years. And now in a world of ever growing ecommerce purchasing lifestyles, is this the way of the future?
Mark Zuckerberg announced on a Facebook Live on May 19, a snap-and-shop feature that allows users to discover, buy and sell items across all of their Facebook platforms. The tool is a “universal product recognition model” that uses artificial intelligence (AI) to identify consumer goods – from fashion to furniture to fast cars. Facebook’s new AI tool will automatically identify items you put up for sale. Facebook’s goal is to create a “Social-First” shopping experience. Product recognition is the first in some of the AI-powered updates coming to Facebook’s ecommerce platform in the near future. Eventually, these will combine not only AI, but also augmented reality (AR), and even digital assistants to develop and create a social-first shopping experience. This auto-tagging feature in Marketplace (coming soon to Facebook Pages) which now autofills all product descriptions during the uploading process, making it easier for businesses to rank in search, while also being able to also communicate seamlessly across multiple different languages. Facebook is aiming to make all photos shoppable across its app, and in time, videos too.
What is GrokNet?
Facebook’s new product recognition AI tool, GrokNet, can identify tens of thousands of different attributes in an image. It is being launched across their marketplace. Mainly beginning by focusing in the Fashion industry. This “AI fashion stylist” could help to offer users personalised shopping experiences and recommendations based on their current wardrobe, different daily suggestions for outfits that can be tailored to the weather and climate. These AI tool attributes will be able to range from specific brands to colours and sizes. Facebook says what makes its tool different is the scope and accuracy.
The Future of Facebook Marketplace
Facebook is also testing versions of this tool that will be built for different businesses. Where they will be able to upload photos onto their business pages containing their own different products, the AI tool will then automatically tag them and link to the certain shopping pages.
While building and developing these tools, Facebook has used its access to users’ photos on Marketplace, to gather data and develop the recognition skills it needs. GrokNet has been trained on a wide database of around 100 million different images, with the majority of these images being taken from Facebook’s current Marketplace. Facebook says this data is vital in creating an AI machine vision system that can identify products in challenging lighting and from dodgy angles — this is still a part of online shopping experience that currently will still not be going away.
Though it is still unclear exactly how accurate GrokNet actually is. Facebook has said it can identify up to 90% of images on Marketplace in the Home, Garden and Fashion category, but has not given similar statistics for other types of product categories yet. As it is often the case with tools such as these, that the difference between the advertised features and actual user experience can be vastly different, and we’ll have to wait and see what reaction GrokNet gets from Facebook’s users.But with nearly 2.7 billion monthly active users and much of the world still under some form of coronavirus restrictions, the push to go big with AI-powered social commerce might just be a silver lining for Facebook and its community. So is this technology truly the future of ecommerce?