In the future, pre-owned products may be AI-certified, which will make it easier for you, as a customer, to trust that the price is reasonable in relation to the product’s wear. The certification, generated using applied AI, will make purchasing pre-owned goods more appealing, which will benefit both the environment and your wallet.
The Swedish waste management association Avfall Sverige recently conducted a study which shows that the reuse of clothing saves ten kilograms more carbon dioxide compared to recycling, and that a mobile phone reused by someone else instead of being recycled saves 60 kilograms of carbon dioxide, which is equivalent to a car driving 460 kilometres. Second hand products and their reuse will become increasingly important for us to reduce our carbon dioxide emissions, and a crucial piece of the puzzle when it comes to tackling climate change. But many people would rather purchase new products than pre-owned.
“One of the reasons may be that pre-owned is uncertain and the pricing unreliable,” says Olof Mogren, Machine Learning researcher at RISE.
“It can be difficult to know whether the price of a pre-owned product is reasonable and that we are paying the true value of the product.”
Easier for consumers to trust the price
This is where artificial intelligence comes in. One way to make pricing more accurate, thereby making it more appealing to buy pre-owned products, is by letting AI decide the price. In a new project at RISE, machine learning will be used to teach computers what pre-owned products are worth. The idea is that, further down the line, second hand stores will utilise the concept to help them set prices on their goods.
“It will be easier for consumers to trust the prices. The pre-owned market will become more stable, which will hopefully increase the incentive to purchase more second hand goods,” explains Mogren.
It will be easier for consumers to trust the prices
AI studies ads and creates own pricing models
In the project, the researchers will teach machine learning models to recognise and understand advertisements for pre-owned products. This means that a computer will view and analyse countless images, descriptions and prices in order to learn what different products cost, how their ads are written, what the associated images look like, and the degree of wear on the products.
“The advantage of letting AI study the ads is that it is so much faster. It would be difficult for us humans to go through so much data. As more and more ads are studied, the models also become more reliable, which makes it easier for people to purchase second hand goods when you know that the price corresponds to the market and is not unreasonably high,” says Mogren.
In the future, stores will be able to use this to set their prices.
“Maybe we could develop an app as a continuation of this project. The stores scan the products with a mobile phone or tablet camera, and the app then suggests a price. It will serve as a certification of sorts. Customers will be able to see that the price is ‘AI-certified’ – a price they know they can trust.”
Not solely for private consumers
The project can also be of benefit to banks and insurance companies, for example. Banks can use machine learning to find out the value of products and generate willingness to invest in circular business models, whereas insurance companies can use AI to help valuate possessions and obtain more accurate valuations.