Manufacturers Face Quality Challenge as AI Moves Into Production and Inspection

Manufacturers are being urged to ensure that quality controls keep pace with the growing adoption of artificial intelligence, as businesses begin moving AI beyond administrative tasks and into production, inspection and other operational processes.

Research from Make UK shows that artificial intelligence remains at an early stage across much of British manufacturing, but expectations for its use are increasing rapidly. Fewer than 40% of manufacturers currently use AI in parts of their businesses and only a small proportion have embedded it widely across their operations.

Quality control remains one of the least developed areas. Make UK found that just 6% of manufacturers currently use AI within quality control, compared with significantly higher adoption in functions such as HR, finance and administration.

The relatively low figure is expected to change as manufacturers explore technologies including automated visual inspection, machine-learning-based defect detection and systems capable of identifying production abnormalities before defective products reach customers.

The potential benefits are considerable. AI-assisted inspection can analyse large quantities of production information and identify patterns that may be difficult for human inspectors to detect consistently. Manufacturers are also exploring systems capable of predicting when processes are beginning to move outside acceptable tolerances, potentially allowing corrective action before nonconforming products are produced.

However, quality professionals are warning that introducing AI into inspection and decision-making also creates new questions about how organisations verify that the technology itself is producing reliable results.

An automated inspection system that incorrectly classifies a defect could allow a nonconforming product to continue through production. Conversely, a system producing excessive false positives could unnecessarily reject acceptable products, increasing scrap, rework and production costs.

The issue becomes particularly important where AI systems continue learning or are updated using new data. Changes to models, training information or operating conditions can potentially affect performance even where the underlying manufacturing process has not changed.

Skills are emerging as another major obstacle. Make UK research found that more than half of manufacturers identify skills shortages as a barrier to greater AI adoption, while manufacturers increasingly require employees capable of working with automation, digital manufacturing and data-driven technologies.

For quality departments, this creates an additional competence challenge. Employees may need to understand not only the product or manufacturing process they are inspecting, but also the limitations of the technology being used to support their decisions.

Industry specialists argue that human oversight will therefore remain important even as inspection becomes increasingly automated. Rather than replacing quality professionals, AI is expected to change their role, with greater emphasis placed on analysing results, investigating unusual patterns and verifying that automated systems continue to operate as intended.

The development is also placing greater emphasis on data quality. Artificial intelligence systems depend heavily on the information used to train and operate them, meaning inaccurate, incomplete or unrepresentative data can directly affect the reliability of their decisions.

This creates a new dimension to traditional quality management. Manufacturers may increasingly need to control datasets, software configurations and algorithm changes with similar discipline to that already applied to specifications, measuring equipment and production processes.

The shift comes as UK manufacturers face wider pressure to improve productivity while managing rising operating costs and persistent shortages of skilled workers. Make UK says manufacturers recognise significant potential for AI to improve productivity and resilience, but adoption remains uneven and heavily concentrated outside core production activities.

As the technology moves closer to the factory floor, quality professionals are expected to play an increasingly important role in determining how automated decisions are validated and monitored.

For businesses operating quality management systems such as ISO 9001, the development raises familiar questions in a new technological setting: whether processes are controlled, whether people are competent, whether changes are properly managed and whether organisations can demonstrate that outputs consistently meet requirements.

AI may ultimately provide manufacturers with some of the most powerful quality-control tools available. The challenge for businesses will be ensuring that the systems used to identify quality problems do not themselves become an uncontrolled source of quality risk.

A message from our sponsors, The Ideas Distillery: 

If you would like to look at how to implement an ISO 9001 quality management system, then simply contact us.

Or, if you want to see what's involved in more detail, then get a completely free, no obligation, totally tailored ISO Gap Analysis for your business (only available to UK businesses).