More than 5,000 French communes use AI to identify poor rubbish sorting

More than 5,000 communes across France now use AI-assisted systems to identify residents who are not sorting their rubbish correctly, in a bid to enforce waste disposal and recycling rules.

Poorly-sorted rubbish can cost local authorities millions as it has to be re-sorted and transported to the correct place once it has been identified.

“If we sort better, our plant will recycle more products and sell more to manufacturers, who will give us more money. So we will send less to landfill. Ultimately, the price we pay for rubbish bins and the household waste tax will be lower or optimised,” said Vincent Chauvet, mayor of Autun, in Saône-et-Loire, to TF1.

If it is found to have incorrectly sorted rubbish, a household will first receive a warning. From the next offence, it will be fined €135.

Around 10 fines for incorrect waste sorting have been issued in Autun since the beginning of the year, Mr Chauvet said. The commune still has an employee whose job it is to check and investigate all waste, and ensure that household rubbish is being sorted properly.

Employees look for identifying details, such as a receipt or bill to identify whose rubbish it is, follow up, and either issue warnings or fines.

Read also: How does recycling work in France?
Read also: French households discard more rubbish despite ‘incentive’ bin fees 

AI-assisted methods

Yet, increased numbers of communes are using AI-assisted methods to identify issues even faster and more accurately. More than 5,000 communes are now using these systems to improve waste collection results.

For example, Reims (Marne, Grand Est) is using small cameras that film the rubbish that comes down into the sorting centre. This is then analysed by AI-assisted technology to identify the waste and geolocate its origin.

Stéphan Rigot, of waste management firm Suez France, said: “The camera films what comes down. And we use artificial intelligence to identify all the faults and sorting errors.” 

This might include oversized cardboard boxes and plastic bags. The system can identify up to 15 types of material.

Door-to-door prevention and education

The data appears in near-real time, and can also help operators to gain more useful insights into the type of rubbish that people tend to sort incorrectly, with the same errors often repeated.

“This clearly shows that communication is needed to help people understand what happened, why they made a particular mistake and why it keeps happening,” said Pascal Lorin, vice president of waste collection for Greater Reims. 

“And that's when we really have to go door to door to explain how simple it is.”

Fly tipping and illegal waste measures

It comes after some French communes have already turned to cameras and AI to detect illegal fly tipping and poor waste sorting at communal rubbish bins, as well as detect the contents of bin bags.

Read also: French AI system can see inside bin bags: what is it being used for? 

Mayors have long been stepping up measures against illegal waste dumping, with many turning to social media to trace and ‘shame’ those who dump waste in nature or in banned areas.