Large language models are showing their potential to transform industrial automation. According to Boston Consulting Group, generative artificial intelligence will surely increase productivity by making people and the machines work together; and there is no place where people and machines interact more than in the industrial automation segments. Aurelien Le Sant, Industrial Automation Technology Director at Schneider Electric, explains more.

These models are created using deep learning algorithms that are trained on large amounts of text data. Recently, they have expanded to various industries and are increasingly applied in industrial automation. The ability of these models to understand and generate human language can greatly improve the efficiency of industrial processes. Our conservative estimates show the potential for these models to eliminate 20% of the effort required by an OEM to build a machine PLC application.

The transition from GPT-2 to GPT-4 allows models to handle more content as inputs, exponentially increasing the applications of large language models in industrial automation. Customizing models for specific use cases or tasks within industrial automation can include code generation, natural language interfaces, and automation system design and development.

The generation of code for control systems such as PLCs or the generation of HMI screens using natural language inputs are some of the possible applications. This reduces the time and effort required to develop control applications. Additionally, these models have the potential to improve the quality of the generated code, leading to fewer errors and faster launch times.

Another application can be the automatic generation of recipe code, which would save time when changing parameters, suppliers or ingredients. The time to create the recipe often affects production time, so any savings here would increase efficiency. Additionally, models can automatically generate documentation associated with the code, such as automated test scripts, which has always been time consuming for the operator.

Large language models can also be used to create natural language interfaces for industrial automation systems, allowing operators to interact with these systems in human language rather than programming languages. This capability allows operators to quickly access existing documentation using natural language commands.

A commonly reported challenge in industrial companies is that domain knowledge may reside with specific people. If that same operator expert knowledge can be fed into secure language models, operators could use voice commands to troubleshoot and take corrective action.

The ethical and responsible use of these models is crucial, considering aspects such as security, data privacy, biases, security and explainability. Furthermore, it is essential to remember that despite the capabilities of large language models, human interaction is still essential and the models must complement these capabilities.

To take full advantage of the benefits of these models, they must be deployed appropriately, taking into account all ethical considerations and limitations. Its use is increasingly widespread across a wide range of industries, changing work practices in industrial automation and significantly improving efficiency and accuracy across the entire work lifecycle, from design and construction to operation and maintenance.