Published on 03-Sep-2026
Looking Back, Looking Forward: What Will NDT Look Like in Five Years?
NDT Talks Episode 7 — The heads of ASNT, ISNT, NDTSS and the ICNDT on what actually changed in the last five years, and the one AI risk none of the optimism accounts for.
Table of Contents
- Session Overview
- Meet the Panel
- The Real Shift: Acceptance, Not Invention
- The Workforce Changed More Than the Technology
- Regional Realities: Four Markets, Four Constraints
- What AI Will Actually Do
- The Risk Nobody Is Planning For: Synthetic Inspection Data
- Rethinking How We Qualify and Certify
- Audience Questions
- Key Takeaways
- Frequently Asked Questions
1. Session Overview
This episode assembled an unusual panel: the sitting presidents of three national and regional NDT societies, the chairman of the international committee that connects them, and a senior operator-side consultant from the Gulf. Between them they hold a view of the industry that few individuals get — across the United States, India, Southeast Asia and the Middle East simultaneously.
The question was deliberately open: what changed in the last five years, and what happens in the next five? The answers converged on something more uncomfortable than a technology forecast. Every panelist named people, not equipment, as the binding constraint — and the session closed on a warning about AI that had nothing to do with job losses.
NDT Talks Episode 7 — Looking Back, Looking Forward: What Will NDT Look Like in 5 Years, in association with NDE 2025, Mumbai:
2. Meet the Panel
3. The Real Shift: Acceptance, Not Invention
Asked to name the single biggest turning point of the last five years, Prof. Balasubramanian made a distinction that framed the rest of the session. The technology is not new. The willingness to use it is.
Prof. Krishnan Balasubramanian — President, ISNT
“Digital technology in NDT has been around for many decades — going back to the early 1980s, digitisation of data was not uncommon. But if you ask what the real turning point is, I would say it is the mindset and the acceptance that digital technology, particularly quantitative digital technology — whether it is radiography, ultrasound, eddy current, and today even magnetic particle and liquid penetrant — has buy-in from the industry towards automation, both for collecting the data automatically and for interpreting it in a different way.”
The End of the High-Fidelity Assumption
His second point is the more technically consequential one, and it inverts an assumption that has driven equipment purchasing for decades.
Prof. Krishnan Balasubramanian — President, ISNT
“Previously we used to go after what I call high fidelity in the data, where high resolution and high accuracy were very important. Today, with digital technology, you can avoid the cost of going for high fidelity data, given that digital data can be processed. There is a significant amount of effort today moving towards data processing and interpretation — whether you use AI or rule-based algorithms is up to you. We are able to do so much better with not so high fidelity data, and sometimes even better than high fidelity data could do for us. The acceptance of digital technology by the last mile operator is probably the biggest shift I see today.”
From Inspection and Reporting to Integrated Data
Dr. Sajeesh Kumar Babu located the change in what happens to the data after the inspection ends.
Dr. Sajeesh Kumar Babu — Chairman, ICNDT and President, NDTSS
“The concept of NDT 4.0 was once more of a theoretical concept. It has now progressed rapidly as a concept of practice, including all those components — AI, digital twins, cloud, IoT, analytics — that are now shaping how inspections are designed and executed. We used to do NDT purely as inspection and reporting. In the last five years you could see a data-driven integrated approach, where the information is fed directly into life cycle management and maintenance planning. It has also helped the process industry move towards predictive integrity management.”
Mr. Abufour agreed from the operator side, with a qualifier the panel returned to repeatedly: two things changed, the way inspections are performed and the way data is analysed, and both were accelerated by the shift to remote working. But in his assessment the industry is still at the beginning, with many sectors requiring substantial development before the gains are broadly available.
4. The Workforce Changed More Than the Technology
Mr. May's answer diverged from the others, and it reframed the discussion for the remainder of the session.
Mr. Clyde May — President, ASNT
“Probably the biggest change I have really seen in the last five years is not so much the use of technology, but that the actual NDT workforce has changed. The acceptance of new technology from NDT technicians is certainly gaining speed and gaining momentum. We have to have that buy-in from the practitioners themselves. It does not matter what technology or what new bells and whistles we bring to the industry — if the technicians do not feel comfortable using them, they are just bells and whistles that sit on the shelf and do not get used.”
He also flagged where adoption has not reached. The gains, in his view, have been concentrated in mechanical integrity and in-service inspection, while construction and fabrication codes and standards remain largely untouched by digital methods.
The Specialisation Problem
Mr. Clyde May — President, ASNT
“The NDT workforce has become very specialised. When we say a UT Level II, you will always see a parenthesis — phased array, or TOFD — they only do one thing. I think we have to be a little careful with that as we go forward, that we are going to turn the workforce into just specialty applications. We did not see that years ago. Everyone was an NDT technician. You learned multiple things, you could do multiple things. Today's workforce is becoming very singular in its focus. We have got to be careful that we keep a well-rounded NDT workforce, not just a specialised workforce.”
Worth noting the tension here: specialisation is a rational response to the demands of in-service mechanical integrity work, where technicians perform very specific tasks repeatedly and get comfortable with them. The cost only becomes visible later, when the industry needs people who can move between methods.
5. Regional Realities: Four Markets, Four Constraints
The most useful segment of the session was the regional comparison, because the constraints turned out to be different in every market.
United States: Scale Without Adoption
Mr. Clyde May — President, ASNT
“In any aspect, particularly energy, oil and gas, we have a tremendous amount of NDE performed every day, and you would be shocked at how much is still even paper driven. Thickness readings being manually recorded. It is amazing to me, in today's world with the technology that we have, that we have not adopted more advanced technology at scale. We love to say how much we have advanced and moved forward, but when you show up on the job sites and look around, it still is very old school.”
The driver, in his analysis, is pricing. At the scale NDT is performed in the US, the commercial question becomes how inexpensively the work can be performed rather than whether it is the most appropriate method — even where advanced technology would cost more today and return that many times over across five years. His conclusion was blunt: the industry may be telling itself it has moved into NDT 4.0 when it has not.
Middle East: Careful, Validated Adoption
Mr. Mohammed Abufour — Global consultant, advanced NDT and inspection technologies
“In oil and gas they are moving towards digitalisation, but they are really very careful. They look at it case by case. Most of the companies here in the region, especially oil and gas, are looking for safety and quality, and they have a lot of challenges they want resolved. So they are moving carefully to get the new technology that helps them — although it is expensive and it is not available to everybody. That is why the approval is not easy to get, and the validation as well. The developer may not like it, but this is the way we have to take it.”
Mr. Mohammed Abufour — Global consultant, advanced NDT and inspection technologies
“We focus on this technology being operated in-house by individuals in the company, because we need to continue it later on. And we have our cyber security — that is also a big problem when we talk about integrating with our facilities. Then the cost, and also the development. Sometimes what we have today we thought was the best, and later on, in six months or a year, something else is developed. We found this with digital radiography — the industry is changing really very fast. So we make sure our selection lands on the right developer.”
An underrated constraint: obsolescence risk. When the technology cycle runs faster than the validation cycle, an operator can complete a rigorous approval process only to find the approved system superseded. That is a procurement problem, not a technical one, and it slows adoption independently of any scepticism about the method.
India: Growth Outrunning the Personnel Pipeline
Prof. Balasubramanian's answer separated two things that are usually conflated. Technology adoption in India is not a national characteristic, it is sector-specific — oil and gas adapts as fast as anywhere in the world, while other sectors sit twenty to forty years behind. The binding constraint is elsewhere entirely.
Prof. Krishnan Balasubramanian — President, ISNT
“India is marching towards becoming a top three economy, and particularly in manufacturing. We will almost double our manufacturing in the next four to five years at most, maybe even earlier. And I am talking about manufacturing across the board — not just shop floors, but oil and gas, petrochemicals, fertiliser, everything put together. Whenever you have manufacturing you need NDT experts, you need NDT personnel to manage this growth. So if we grow by 100% over the next four years, that means we need to at least double our certified NDT personnel. That is where the challenge lies.”
Prof. Krishnan Balasubramanian — President, ISNT
“India does produce its share of NDT engineers and NDT technicians, and that is a big benefit of being in India. However, the biggest problem we have is due to salary disparity. We lose all the good people. We lose all the good people to other countries, and hence we are playing catch-up. I have nothing against these people finding a better salary for themselves, but qualifying personnel and retaining qualified personnel is a major challenge we are facing today, and will probably continue to face.”
He was careful not to present this as a problem with an easy technological fix. Robotics and AI may augment capacity in certain areas, but how far that augmentation extends remains an open question, and the questions he posed back to industry were about pay and about making the work feel more challenging — not about tooling.
Asia-Pacific: Growth Without Harmonisation
Dr. Sajeesh Kumar Babu — Chairman, ICNDT and President, NDTSS
“Asia-Pacific is the fastest growing NDT market, driven by mega infrastructure growth, energy demands, transportation and manufacturing expansion in both India and China. However, what makes us different from the West is that Asia-Pacific is not harmonised, especially in the NDT regulatory frameworks. We have wide variations between countries. Some economies enforce strong statutory NDT rules — Japan, Korea, Australia — while others run on project-based specifications. Civil engineering NDT has almost no regional standardisation, creating inconsistent acceptance by asset owners.”
The consequences he listed are commercial: delays, rework, inconsistent quality, limited cross-border labour mobility and barriers for multinational operators. Personnel certification is similarly uneven — strong in some countries, mixed across ASEAN and South Asia.
Dr. Sajeesh Kumar Babu — Chairman, ICNDT and President, NDTSS
“The biggest bottleneck is the ageing workforce, or the shift of the workforce. The Level III workforce has been reducing, or it is not meeting the demand, and there is a low inflow of young professionals — they see NDT as a high risk field with a slow career path. We also have limited training capacity in advanced NDT methods including phased array and digital radiography, and there is no established ecosystem yet for AI-enabled NDT or digital twin networks.”
The Four Markets Compared
6. What AI Will Actually Do
The moderator put the question directly: is AI a threat, an augmentation, or a replacement? No panelist argued for replacement, but their reasons differed usefully.
Dr. Sajeesh Kumar Babu — Chairman, ICNDT and President, NDTSS
“It will not be a complete shift where AI replaces. What you will see is the technology moving from seeing defects to understanding behaviour. In the next five years or a decade, the routine radiographs, thickness scans or composite inspections will be pre-screened by validated AI tools — and AI tools need to be validated as well when you put them into practice. The human experts, engineers and highly skilled technicians will focus on the edge cases, where the pass-fail criteria are crucial, on correlations and acceptance decisions, rather than on the first level of defect hunting.”
Digital Twins for Ageing Infrastructure
Dr. Sajeesh Kumar Babu — Chairman, ICNDT and President, NDTSS
“In my area, in Southeast Asia, we have large scale assets — very expensive bridges, cable-stayed bridges, infrastructure, underground tunnels dating back to the 1970s or even earlier. These infrastructures are ageing, and you will see a lot of digital twin based NDT and structural health monitoring moving in line together to consolidate data. It becomes a living digital twin where we can see the changes in the infrastructure, and it helps in residual life prediction, so you can replace a structure much earlier rather than leading to a disaster.”
Permanently installed monitoring is what makes the living digital twin possible, see our full guide to continuous monitoring in NDT.
→ Read: IoT and NDT for Monitoring and Predictive Maintenance
Human-Led, and a Warning About the Vocabulary
Mr. Clyde May — President, ASNT
“I do not believe AI is going to replace humans. I believe AI will be a human-led tool, not the reverse. The demand for NDT is going to continue to grow as assets and infrastructure age — there is no question, we are not going to see a downturn in the need for NDT. So how do we close the gaps? That is probably the first and most important place for us to look at where we can use AI to augment — not necessarily to replace the human, but to allow the human to be more efficient and more focused on the things that require human intervention.”
Mr. Clyde May — President, ASNT
“We hear the word AI so much, and everything is AI enabled. I think we have to be very careful when we use that term — it certainly gets overused. We hear things like ADR, automated defect recognition. We really need to stop using terms like that, because a defect is something that is unacceptable. We are looking for discontinuities, and then we decide whether they are a defect or not.”
That terminology point is not pedantry: if a tool is described as recognising defects, it implies an acceptance decision has already been made. Calling it what it is — discontinuity detection — keeps the adjudication where it belongs, with the qualified person.
Smart Hardware Is Available Now and Barely Used
Mr. Clyde May — President, ASNT
“I am going to use the term smart hardware and smart software rather than AI enabled, because as soon as you say AI, we kind of lose exactly what it is we are talking about. One of the challenges, as an example, for the implementation of phased array for weld inspection is that there is a tremendous amount of variables and a tremendous amount of knowledge currently needed boots on the ground at that weld. But if we can do some of that remotely — that is smart hardware. We can dial into the equipment. If we have a bevel angle change and we need to rework our scan plan, rather than having someone physically on the ground doing that, we could do that easily remotely today. We just are not adopting it. Is that AI? I would not call it AI, but it is smart hardware.”
The Operator's View: AI as a Calculator
Mr. Mohammed Abufour — Global consultant, advanced NDT and inspection technologies
“AI is a tool that helps me, and I need it in many aspects, especially with ageing infrastructure where I want to make sure it is safe and there is no defect present. In construction we do thousands of film interpretations. I have worked for 40 years, and we face individual missed calls or over calls. AI would help me tell where to focus, and that is what we need — the speed. If the standard asks me for five or ten per cent validation, I can do more than that using AI, and I can also verify the quality of the film. It is like a calculator: it multiplies a big number by a big number and gives me the answer, but led by me. I am the responsible one.”
He extended the point to remaining life calculation, where a result that once took manual effort can now be produced quickly and sent to a colleague in another country for a second opinion within the same working day. But the accountability does not move. As he put it, if there is a failure later, he is the one who is responsible — which is why the tool operates under his rules and his guidance.
AI-assisted film and image interpretation is the most mature application discussed here, see our deeper look at how it is being applied.
→ Read: AI Digital Radiography Interpretation
7. The Risk Nobody Is Planning For: Synthetic Inspection Data
The session's most important contribution came at the end of the AI discussion, when Prof. Balasubramanian separated two questions that are usually asked as one.
Prof. Krishnan Balasubramanian — President, ISNT
“Should we be scared of AI today? The answer is no, because today AI has some limitations, as was put forth by all the speakers before me. It is going to help you do a lot of chore work. Then the question that comes up is, should we be afraid of AI? And the answer is yes. Just like we have fake videos and fake images today, for me to create a digital X-ray image, or to create ultrasonic phased array data that is not real but very difficult to identify as such, is possible — and it will be generated by AI. Previously we used to see data that was photoshopped, easy to pick up. Now more and more we are seeing data generated by AI.”
Prof. Krishnan Balasubramanian — President, ISNT
“This is where the worry comes in. You have fake data, which means you have fake billing, and you also have a lot of welds and other structures that are not being inspected but are being certified as inspected. These will actually go against what we all stand for. I am not too worried about AI taking away the job — right from the time computers were built we have always managed to morph ourselves and find better and higher paying jobs. But I am worried that data manipulation and fake data are going to be a concern.”
Why this deserves more attention than it gets: every other AI discussion in NDT concerns whether the technology is good enough to be trusted. This one concerns whether the data reaching the reviewer is real at all. The controls for it — provenance, chain of custody, tamper-evident acquisition, audit of raw data rather than reports — are largely absent from current practice, and they are not the same controls that validate an algorithm.
8. Rethinking How We Qualify and Certify
If the workforce is the constraint, the qualification system is where the panel looked for movement.
Mr. Clyde May — President, ASNT
“At some point we are going to have to look at the 60-year-old approach of saying this is how we qualify and certify individuals. What we were doing 60 years ago does not fit today's workforce. My career began almost 50 years ago, and it was a career doing multiple things. That is not done today. Individuals focus on very specific things — they are very task focused now. So at some point we are going to have to start training to task, and qualifying and certifying people to do tasks. The industry changed, and we are still using models from 60 years ago.”
His argument for change is partly generational. Fewer entrants are willing to spend two and a half years training and certifying to do something they could be productive at within six months if trained to the specific task. Set against his own earlier warning about over-specialisation, that is a genuine tension the industry has not resolved.
Employer-Based Certification and the Integrity Question
Dr. Sajeesh Kumar Babu — Chairman, ICNDT and President, NDTSS
“It is not only Asia, it is also the rest of the world. There is always a challenge in delivering employer-based certification in a proper manner, and it is on both aspects — the understanding of employer-based certification, and the integrity of managing it. That is why most countries have understood this problem and developed international standards, such as ISO 9712 in general, or in aerospace where it is controlled through the National Aerospace NDT Board framework. In a fast-paced business, employers try to get these certifications quickly done without the necessary competence.”
His view is that accreditation is now rebalancing this. Where NDT operates under ISO/IEC 17025 as a laboratory or ISO/IEC 17020 as an inspection body, national regulators enter the picture. The model he expects to spread is a central certification functioning like a driving licence, with employer certification retained on top of it for specific employer needs — the arrangement already common in aerospace.
The Operator as Enforcement Mechanism
Mr. Mohammed Abufour — Global consultant, advanced NDT and inspection technologies
“For employer certification, in most of the oil and gas companies here we have a qualification test. That test pushes the service provider or the employer to upgrade their qualification before they come to our test, because if they come and fail, they have to come back and sit again. Since he is looking for a job or for our services, he will upgrade his employer certificate to meet the oil and gas requirements. So we are very strict with the qualification exam. If he fails, he will not come the next day — he will come after a month, then three months, then six months.”
The mechanism is worth isolating: a rigorous client-side qualification test transfers the cost of weak employer certification back onto the employer. The commercial penalty for sending an underqualified technician does the enforcement that the certification scheme itself cannot.
9. Audience Questions
What is the most valuable skill an inspector can offer in the future?
Mr. Clyde May — President, ASNT
“Become as well-rounded as you can possibly be. Learn as many things as you possibly can, and try not to focus on a singular thing. But probably the most valuable skill is the ability to reason and think. We rely too much on apps, we rely too much on that instant answer rather than problem solving, and I think we are losing our ability to problem solve. For a technician in the field, what we go to inspect is rarely what we find when we arrive, or the conditions change. How do you adapt? How do you problem solve in the moment? If I am looking to hire someone, that is one of the things I look for.”
Read alongside the AI discussion, that answer is consistent rather than nostalgic. If validated tools pre-screen the routine work and humans handle the edge cases, then adaptive judgement is precisely the capability that increases in value — and it is the one the panel felt is least deliberately taught.
10. Key Takeaways
- The turning point of the last five years was acceptance, not invention. Digital NDT data has existed since the 1980s; what changed is buy-in from industry and from last mile operators.
- High fidelity data is no longer the default goal. Where data can be processed, lower fidelity acquisition plus strong interpretation can outperform expensive high-resolution capture.
- NDT 4.0 moved from concept to practice, with inspection output now feeding life cycle management and maintenance planning rather than terminating in a report.
- Digital adoption is concentrated in in-service mechanical integrity. Construction and fabrication codes and standards remain largely untouched.
- The workforce, not the technology, is the binding constraint in every region represented on the panel.
- Over-specialisation is a growing risk. A UT Level II qualified only in one technique is a rational response to in-service work, but it erodes the well-rounded technician base.
- Regional constraints differ: the US lags on adoption despite scale, the Middle East is slowed by validation cycles and obsolescence risk, India must double certified personnel while losing people to higher salaries abroad, and Asia-Pacific lacks regulatory harmonisation.
- No panelist expects AI to replace inspectors. The expected model is validated AI pre-screening routine work while qualified people handle edge cases and acceptance decisions.
- Terminology matters. Automated defect recognition is a misnomer — tools detect discontinuities, and the decision on whether a discontinuity is a defect stays with the qualified person.
- Smart hardware, such as remote access to phased array equipment for scan plan changes, is available today and barely adopted. It closes competency gaps without any AI involvement.
- The most serious AI risk raised was synthetic inspection data — realistic but fabricated radiographs and phased array datasets enabling fake billing and structures certified as inspected without inspection.
- Qualification models are 60 years old and fit poorly with a task-focused workforce. Training and certifying to a task is one proposed direction, in tension with the warning against over-specialisation.
11. Frequently Asked Questions
What is NDT 4.0?
NDT 4.0 describes the integration of non-destructive testing with digital industrial technologies — artificial intelligence, digital twins, cloud infrastructure, the industrial internet of things and data analytics. The panel's position was that it has moved from a largely theoretical concept to one that shapes how inspections are designed and executed in practice, with inspection data feeding directly into asset life cycle management and maintenance planning rather than ending in a standalone report.
Will AI replace NDT inspectors?
No panelist expected replacement. The consistent view was that AI functions as a human-led tool. The model described is validated AI pre-screening routine work such as radiographs, thickness scans and composite inspections, with qualified people concentrating on edge cases, correlations and acceptance decisions. Accountability for the result remains with the person who signs it, which is the practical reason a human stays in the loop regardless of how capable the tool becomes.
What is the biggest risk of using AI in NDT?
The risk highlighted in this session was synthetic inspection data — realistic digital radiographs or ultrasonic phased array datasets generated by AI rather than acquired from a real component. Unlike earlier image manipulation, which was comparatively easy to identify, AI-generated data can be difficult to distinguish from genuine acquisition. The consequences described were fraudulent billing and welds or structures certified as inspected when no inspection took place.
Why is there an NDT workforce shortage?
Several causes were identified. Demand is rising as assets and infrastructure age. The existing Level III population is aging and shrinking. Fewer young professionals are entering, with NDT perceived as a high risk field with a slow career path. Training capacity in advanced methods such as phased array and digital radiography is limited. And in some countries, including India, qualified personnel leave for higher salaries elsewhere, so the pipeline drains as it fills.
What is the difference between central and employer-based NDT certification?
Central or third-party certification is issued by an independent body against an international standard such as ISO 9712, and it travels with the individual. Employer-based certification is issued by the employer against its own written practice and is valid within that organisation. The panel's concern with employer-based schemes was the integrity of their administration under commercial time pressure, and the expected direction is a central certification functioning as a baseline with employer certification retained on top for specific needs.
How are digital twins used in NDT?
Digital twins are being combined with structural health monitoring on large ageing assets such as bridges, tunnels and major infrastructure. Inspection and monitoring data is consolidated into a model that reflects the current condition of the structure rather than a snapshot from the last inspection. The practical value described is residual life prediction, allowing a structure to be repaired or replaced on evidence rather than after a failure.
Ads
