AI as a tool for certification
02 June 2023, 11:19
The need for certification of products and services is predicted to increase in the future. RISE, which is both a certification authority and has research expertise in AI, has looked at opportunities to increase customer value in certification processes using artificial intelligence.
There will be an increasing need for certifications in the coming years. There are several reasons for this - an increasing need for quality assurance of products and services, new regulations from the EU, and safety and sustainability aspects. It can concern products in sports, construction, electronics, security, transport, and more. A third reviewing partner is important - RISE is one such partner.
A certification process works in different ways depending on the product's risk. The level of control can vary. In the case of very low risk, companies can analyse their impact themselves based on standards and use this to achieve the certification target, but in the case of products or services with higher risks, tests are needed in testing facilities via an accredited third part. These tests usually result in a report. In the case of large-scale production of products, it is also important to control the production to ensure that all products are of the same standard, whether it is the first or the millionth product to be checked. To become certified, companies need to submit data, inspection certificates and technical documentation to a certification authority or notified body (CE marking), after which they receive a certificate or the right to CE mark their product. It may also be relevant to certify people, for example in the construction industry. Certification is a continuous process that needs to be done in a similar way every time.
AI can be a tool in certification processes
RISE is an approved certification authority and owns around 130 test beds, several of which can also be used for testing products, such as penetration tests in cyber security. AI can streamline the work that RISE already does. The certifications are carried out in a systematic way, where you know what you want at the end, a certificate. If AI were to complement humans, we would like the quality of the result to be at least as good as if a human were doing it. Here you can see a value in using different types of AI models, not least the type of large language models that have become popular recently. These can be used to process and summarise text. This can be useful for summarising submitted product texts or checking competencies required for personal certifications. It is also possible to use AI for other types of data than text, such as transcribing and summarising meetings, analysing customer satisfaction via interaction patterns or monitoring a process where you want to assess, for example, abrasions in a product.
The value of using AI is to get an equivalent or better result, but with fewer resources. In addition, certification organisation can be relieved and have more time for personal customer dialogue and increased customer value. There are three main elements to be achieved in future certification processes - more efficient processes, increased quality assurance and increased customer value.
Risks and security in AI implementation
Data security is an important consideration when implementing AI in your processes, whether for certification or other applications.
"You can't assume that an AI model is always right, so it's important to set up systems and questions in several stages so that you can be confident of the results. The safety bar must sometimes be set very high, for example, when self-driving cars are to make decisions, while when it comes to simpler tasks, such as counting migratory birds, a certain degree of error tolerance may be acceptable. It is important to standardise the type of response required from the AI model so that you can get the desired solutions," says Peter Nordström, a researcher in data analysis and digitalisation with a focus on AI at the Centre for Applied AI.
RISE is helpful
RISE, which has expertise in AI and the domain, but also coaches to drive development work in companies with the goal of an AI strategy, can be a good partner and catalyst when it comes to AI transformation and applications. By identifying opportunities with AI and identifying where in the processes it is useful, RISE can provide suggestions for efficiency improvements and opportunities with AI.
Read more: Certification at RISE
2024-11-04
2024-09-09
2024-09-03
2024-09-02
2024-08-06
2024-04-30
2024-04-29
2024-03-19
2024-02-06
2023-11-15
2023-10-02
2023-09-13
2023-08-24
2023-06-21
2023-06-19
2023-06-02
2023-05-17
2023-05-09
2023-04-27
2023-04-05
2023-04-04
2023-04-04
2023-03-29
2023-03-16
2023-01-31
2023-01-31
2022-12-06
2022-11-15
2022-10-24
2022-10-24
2022-10-20
2022-10-20