A quartet of thoughts

On the day after the Prime Minister extolled the virtues of artificial intelligence (AI), and how it would help the country grow and increase productivity, I had a conversation that caused me to worry about the future. The individual concerned had had to book an appointment and the only way to do this was through a website. The relevant details were input and the appointment confirmed. When they attended for the appointment, they were informed that they did not meet the criteria and would have to go away and arrange an alternative appointment elsewhere. All this was just a waste of everyone’s time and could easily have been avoided. Nowhere on the website did it warn of the specific criteria, which would have been simple to implement, and nor did it flag up any issue when the relevant details were input and submitted. All of 45 years ago I was using IF THEN logic when programming. Why can’t we implement it today?

I suspect that the programmers of the appointment system did not think through all the requirements of the potential users. If this occurs with less complicated systems, what are the implications for AI? I think the lessons for non-destructive testing (NDT) are:

• Before adopting the most advanced technology, explore simpler solutions and implement any improvements in existing processes.

• Understand the importance of cross-discipline knowledge to ensure successful implementation: knowledge of the NDT application as well as knowledge about the capabilities of the technology.

I attended an excellent talk on uncertainty given by a colleague. In order to make it relevant to a wide audience, who did not necessarily calculate uncertainty estimates as part of their work, and to avoid slides covered in equations, she drew an analogy with an example from everyday communication. The lesson was that, instead of being just a luxury add-on to a statement of measurement, uncertainty estimates enhance the information, allowing the recipient to have more confidence in decision-making based on the measurement. This was reinforced when travelling on the rail network not long afterwards. The station display gives the actual estimated time of arrival, which is useful, as long as they do not keep incrementing it before the train eventually arrives. However, on this day the display suddenly announced ‘Delayed’. So, do you wait? Go away and come back? Look for alternative means of getting to your destination? I understand that transport has delays and it is not always possible to say when the delay will be resolved, but it does illustrate how providing an estimate of uncertainty improves the quality of information for the recipient. In NDT, despite the importance of the decisions made on the back of inspection results, we have been reticent to apply uncertainty estimates. This will take on more importance when AI is implemented. I know there is research work being undertaken in assessing uncertainty in machine learning, but will this be implemented in practical applications?

There is a worrying trend amongst those being questioned, in a media interview or a general meeting, for them to immediately respond with: “That is a good/excellent question”. Questions may be incisive but the use of adjectives such as good and excellent implies that there is a quality hierarchy of questions; that other questions asked are average or poor. The purpose of a question is to elicit information to increase the understanding or knowledge of the individual or a group. If the question achieves this purpose, then the oft stated phrase: “there is no such thing as a stupid question” applies. In NDT, it is particularly important to ask questions if there is any uncertainty or ambiguity. Any lack of awareness and understanding leads to errors and accidents.

I have acquired some new knowledge, which goes to prove that you are never too old to learn 😊. I am sharing this knowledge with you, so that in future you can avoid that awkward silence when someone asks you what you do and you reply: “NDT Engineer” or “NDT Operator”. What you should reply is that you are a Kung Fu NDT Engineer or Kung Fu NDT Operator. I am sure that this will engender further interest and keep the conversation going. I have learnt that the Chinese character for ‘kung’ can mean skilful work, hard training or endeavour, while the character for ‘fu’ means time spent. The combination of the two means time spent at skilful work or hard training. NDT demands both time applying skill and undertaking hard, ongoing training. So, you can legitimately call yourselves Kung Fu NDT Engineers or Kung Fu NDT Operators.

Please note that the views expressed in this column are the author’s own personal ramblings for the purpose of encouraging discussion within NDT News. They do not represent the views of Amentum or BINDT.

Letters can be mailed to The Editor, NDT News, Midsummer House, Riverside Way, Bedford Road, Northampton NN1 5NX, UK. Email: ndtnews@bindt.org or email Bernard McGrath direct at bernard.mcgrath1@global.amentum.com