Amazon announces new machine learning tools to help customers monitor machines and worker safety

AI-enhanced tools promise safety and efficiency gains in industrial settings, but privacy campaigners fear mission creep

Amazon Web Services (AWS) on Tuesday launched five new industrial machine learning services aimed at helping industrial plants and factories to improve safety, operational efficiency, and quality control at their workplace.

The company said that companies can use these services to embed artificial intelligence (AI) in their production processes to identify productivity bottlenecks, potential equipment faults, and worker safety and compliance violations.

The five tools, named Amazon Monitron, AWS Panorama Software Development Kit (SDK), AWS Panorama Appliance, Amazon Lookout for Vision and Amazon Lookout for Equipment, combine computer vision, sensor analysis and machine learning capabilities to address technical challenges faced by industrial customers.

The launching of these new services also indicates Amazon's growing ambitions to strengthen its position as a leading player in the industrial cloud sector.

According to Amazon, its Monitron tool is comprised of a gateway, sensors, and machine learning software. The small sensor in Monitron can be attached to equipment to detect abnormal conditions, such as high or low temperatures or vibrations, and predict potential failures.

AWS says it is already using 1,000 Monitron sensors at its fulfilment centres near Mönchengladbach in Germany to monitor conveyor belts handling packages.

AWS Panorama Appliance, meanwhile, enables industrial facilities to use their existing cameras to improve safety and quality control. The tool uses computer vision to analyse video footage and detect safety and compliance issues.

According to the Financial Times, AWS Panorama can be used to detect vehicles bring driven in places where they are not supposed to be. Some big companies, including Deloitte and Siemens, are already testing the system, it said. AWS Panorama SDK allows industrial camera makers to embed computer vision capabilities in their new cameras.

Amazon Lookout for Vision is designed to find flaws and anomalies in processes or products by utilising AWS-trained computer vision models on videos and images.

Amazon Lookout for Equipment gives customers with existing equipment sensors the ability to use machine learning models to detect unusual equipment behaviour to predict future faults.

While AWS claims that industrial plants can use these new tools to improve productivity and safety at their workplaces, privacy campaigners have also raised concerns about these tools.

Earlier this week, the Trades Union Congress (TUC) in the UK released its report into the impact of AI-powered tools on well-being of workers. The report warned that some intrusive technologies being used in companies can have potentially negative effects on "workers' well-being, right to privacy, data protection rights and the right not be discriminated against".

Silkie Carlo, director of privacy group Big Brother Watch, told the BBC that automated workplace monitoring "rarely results in benefits for employees".

"It's a great shame that social distancing has been leapt on by Amazon as yet another excuse for data collection and surveillance," she added.

With concerns about workplace surveillance rising, this week Microsoft apologised for a new productivity score featured introduced in Microsoft 365, which could be used to track individuals' detailed usage of the cloud based productivity suite by administrators. Microsoft says it will remove individual usernames from the productivity score feature.