The Implications of Artificial Intelligence for the Quality Infrastructure

Artificial intelligence (AI) is on everyone’s lips. A significant stimulus was when the US company OpenAI published the ChatGPT in November 2022.[1] GPT stands for Generative Pre-trained Transformer and is a text-based dialogue system as a user interface based on machine learning. Millions of internet users registered to gain access to this new oracle in just a few days.

Since quality infrastructure professionals are working intensively on digitalisation, we have asked ChatGPT about the significance of artificial intelligence for quality infrastructure. The chatbot generated the answers to the following questions [2]:

QI4D: What are the implications of artificial intelligence for quality?

ChatGPT: The implications of artificial intelligence for quality infrastructure are wide-ranging. AI-driven technologies can help organisations achieve higher quality assurance and control levels, allowing for more automation and better data analysis. AI can also improve the efficiency of quality control processes, such as testing, validation, and inspection. Additionally, AI can help to identify defects in products and systems and provide recommendations for improvement. Finally, AI can be used to inform decision-making when it comes to standards and regulations, helping organisations to maintain compliance with industry requirements.

QI4D: Do you know examples of how AI helped to improve quality?

ChatGPT: Yes, AI has been used to improve quality in various ways. For example, AI-driven technologies can automate quality assurance processes, reducing the time needed to complete tests and inspections. AI can also identify potential sources of errors or defects in products or systems and provide recommendations for improvement. Additionally, AI can be used to inform decision-making when it comes to standards and regulations, helping organisations to maintain compliance with industry requirements. Finally, AI can provide insights into the effectiveness of current standards and regulations, allowing for more informed decisions and better policies.

QI4D: Which countries are leading to the use of AI in quality management?

ChatGPT: Many countries worldwide are leading the way in using AI for quality management. The most prominent examples include the US, China, Japan, Singapore, and South Korea. In the US, many companies are leveraging AI to automate quality assurance processes, reduce the time needed to complete tests and inspections, and improve the accuracy of results. In China, AI-driven technologies are being used to enhance the effectiveness of quality control processes, such as testing, validation, and inspection. In Japan, AI is being used to help organisations maintain compliance with industry standards and regulations. Finally, AI is used in Singapore and South Korea to inform decision-making regarding standards and regulations.

QI4D: What are the risks of using AI in quality management and qual infrastructure?

ChatGPT: Several risks are associated with using AI in quality management and quality infrastructure. For example, AI-driven technologies can be unreliable and produce inaccurate results, leading to errors and defects in products and systems. Additionally, AI can be vulnerable to malicious actors, who can potentially exploit flaws in the system to compromise data or manipulate results. Finally, AI can be difficult to control and regulate, meaning that organisations may struggle to ensure compliance with industry standards and regulations. Therefore, organisations must take appropriate steps to mitigate these risks when leveraging AI for quality management and infrastructure.

ChatGPT’s responses point to significant development trends and implications for quality infrastructure. They highlight the benefits of using AI in the industry, especially in quality assurance and quality control. QI bodies and regulators can also benefit from the use of AI.

ChatGPT also identifies various risks associated with the use of AI. However, the statements remain general and repetitive, even when asked for examples. Here it is important to remember that ChatGPT is still a beta version whose knowledge base is currently at the end of 2021. This means that the system still needs to learn about current utilisation.

The foundations for AI are only being laid today in standardization. The joint technical committee of ISO and IEC JTC 1/SC 42, whose secretariat is provided by the US ANSI, plays an important role here.[3]

SC 42 focuses on JTC standardization programs on AI and guides JTC 1, IEC and ISO committees in developing Artificial Intelligence applications. The committee is developing international standards for AI intending to accelerate AI adoption while simultaneously addressing emerging issues to enable the successful digitalisation of sectors such as smart manufacturing.

SC 42 has developed an AI ecosystem platform approach. It is designed to allow a wide range of stakeholders to address barriers to the adoption and diffusion of AI while simultaneously addressing ethical and societal concerns. The following diagram illustrates the approach:

Source: ISO /IEC JTC 1/SC42. Trustworthiness of IoT Systems. [4]

To support the use of AI at the enterprise level, SC 42 has developed the ISO/IEC 42001 standard for an AI management system.

However, the successful application of AI also requires reliable discoverable, accessible, interoperable, and reusable data (FAIR). The metrology community is challenged to develop qualitative metrics to explain AI methods’ ability and interpretability. The consideration and operationality of measurement uncertainty play a central role here. A new “metrology for data” must be established.[5]

Artificial intelligence is also already being used in conformity assessment. One example is the French company Institut Technique Gaz et Air (ITGA), which is developing a new test method that delegates the detection of asbestos fibres to artificial intelligence. This test is used in air monitoring for public health protection and occupational safety. The accreditation body Cofrac had to develop a new scheme for the accreditation these new services.[6]

Another general example is described in a recent TIC Council publication on remote conformity assessment. It highlights that the software used to perform remote conformity assessment activities may have artificial intelligence technologies for recognising and classifying defects and IoT for collecting parameters and evaluating acceptability and working condition. These technologies support the auditors in their decision-making process. [7]

In the context of market surveillance, Germany is developing artificial intelligence tools to identify potentially dangerous products by scanning online customer reviews. Mounting complaints about the missing safety of a given product can provide a lead for surveilling this product. [8]

ChatGPT’s responses need to answer what AI means for developing and emerging countries. Industries and services will also use these new technologies to compete. However, the question arises to what extent Chat GPT and other AI answers will adapt to the specific realities of the countries of the South when the bulk of the data is generated in the industrialised pre-economies.

Conversely, AI could also represent an opportunity for developing and emerging countries, as it makes standardized knowledge more easily accessible. Quality infrastructure institutions will be challenged to support the domestic economy in harnessing this potential.

References

[1] ChatGPT

[2] OpenAI  (Retrieved 28/01/23)

[3] ISO/IEC JTC 1/SC 42

[4] Diab, W.W./ Ferraro, A./ Klenz, B./ Lin, S.W./ Liongosari, Edy/ Tannous, W.E./ Zarkout, B. (2022); Industrial IoT Artificial Intelligence Framework, IIC Industry IoT Consortium

[5] Eichstädt, S., Keidel, A. & Tesch, J. 2021. Metrology for the digital age. Measurement: Sensors, 18

[6] ILAC (2022) Cofrac – artificial intelligence and asbestos detection, post 20/04/2022

[7] TIC (2022), TIC Council. Horizontal Task Force on Remote Activities. Publication. Remote Activities of Conformity Assessment. First edition. March 2022

[8] BMWK (2022). “United in Quality and Safety – An introduction to quality infrastructure in Germany and the European Union for policymakers and trade partners“, Berlin