Dust hangs thick in the air at Sharp Group’s recycling plant in Rainham, east London, where the relentless rumble of hoppers and conveyor belts sets the grim pace. It is, by any standard, an unforgiving place to make a living, and increasingly that is where the problem lies.
The family-run container and waste management business, which processes up to 280,000 tonnes of mixed recycling annually, relies on 24 temporary workers stationed along its high-speed conveyor belts. They sift through a procession of debris in real time, ranging from old trainers and VHS tapes to concrete slabs. It’s the kind of work few people line up for, and the numbers prove it. Annual staff turnover at the factory is 40%, reflecting an industry-wide retention crisis that is now forcing British SMEs to confront a question once reserved for car factories and Amazon warehouses: Can robots do this instead?
For Sharp Group, the answer could already be taking shape in the line itself. A humanoid robot named Alpha, the Automated Litter Processing Humanoid Assistant, is trained to sift through the waste stream alongside the human collectors it could one day replace. Alpha was built by Chinese company RealMan Robotics and adapted to British recycling conditions by London-based TeknTrash Robotics. It represents an unusual bet on humanoid form factors in an industry previously focused on bespoke automated kits.
“The appeal of a humanoid is that you can put it here and it will stay there,” says Chelsea Sharp, the plant’s finance director and granddaughter of founder Tom Sharp. “It will be picked all day, 24 hours a day, seven days a week. There will be no vacation requests, there will be no sick days.”
This blunt commercial logic is countered by an equally blunt security case. Work-related injuries and illnesses in the waste sector are 45% higher than the national average in other industries, and the mortality rate is a significant multiple of the total workforce. Sharp Group prides itself on its own security record, but recruiting in such an environment is becoming increasingly difficult to defend.
“The line is moving all the time, you’re constantly picking. I have a lot of pickers because they’re just not up to the task,” says line manager Ken Dordoy. The company rotates its employees through different waste streams every 20 minutes and ensures rest with regular breaks – a system that speaks volumes about the stress involved.
Alpha is not a quick fix for now. It’s in the early stages of a comprehensive training program in which a plant worker wears a VR headset next to the robot to demonstrate what a good harvest looks like. The dual challenge, explains TeknTrash founder and CEO Al Costa, is to teach the machine to first identify objects on a conveyor belt and then reliably lift them. His company’s HoloLab system feeds Alpha a deluge of data from multiple cameras, generating millions of training data points every day.
Costa is open about the gap between marketing hype and operational reality. “The market believes these robots are ready-to-wear, that all you have to do is plug them in and they will work perfectly. But they require extensive data to be effectively useful.”
The humanoid approach has the advantage of being able to integrate into existing infrastructure without expensive plant redesign, which is no small consideration for SMEs operating on the low profit margins typical of the recycling sector. The alternative, increasingly preferred by larger operators, is extensive retrofitting with bespoke automated kits.
Colorado-based AMP, which operates three of its own factories and supplies equipment to dozens of facilities across Europe and the UK, is taking this approach. Its systems use air jets to shoot objects into chutes, with AI continually improving the machine’s ability to identify and sort materials. “Our robots are much more efficient than humans, probably eight or ten times faster,” says managing director Tim Stuart. “AI technology and jets have significantly increased the capacity, efficiency and accuracy of what we can do.”
California’s Glacier, co-founded by Rebecca Hu-Thrams, uses mounted robotic arms paired with AI vision. She quickly notices the sheer unpredictability of the material her machines have to contend with. A leaky beer can can endanger sensitive equipment; Her customers, she adds, have seen “unbelievable things like hand grenades and guns coming through her facility.” The proposal, she says, is a large-scale improvement: “As our models learn from more than a billion elements, AI keeps getting better. And we’ve always designed our technology to work not just for large urban facilities, but also for semi-rural facilities that operate on much tighter budgets.”
With all the different approaches, the entire industry comes to a consistent conclusion. The labour-intensive model that has supported Britain’s waste processing for decades is reaching the end of its useful life. Scientists who study the sector see the same development. Professor Marian Chertow of Yale University argues that “robotics combined with AI-driven vision systems offer the greatest potential for improving material recovery, work experience and economic competitiveness in the recycling sector.”
The unpleasant question remains what happens to the people who are currently picking. Chelsea Sharp doesn’t claim that the work is anything but strenuous. “It’s a really dirty place to work. You can see the dust, you can hear the noise. It’s not that nice.” However, their stated plan is more retraining than replacement. “The plan is to upskill these employees. They will maintain and supervise the robots. And it keeps these same people away from all hazards, including the unpleasant environment, heavy lifting and noise.”
Whether the rest of the industry follows Sharp’s example or whether automation automatically leads to a calmer, leaner workforce remains to be seen in the next few years. What is undisputed is that the UK recycling line in 2030 will not be comparable to the one in Rainham today.
