Will AI Take My Job? 7 Reasons Asbestos Surveying Is Hard to Automate
AI can draft a report in seconds. Establishing what is hidden inside an ageing building is a very different challenge.
“Will AI take my job?” has become a serious career question. For anyone considering asbestos surveying or asbestos analysis, there is a good reason to look beyond the headlines: much of the value of these roles comes from work that happens outside a computer.
Consider an older building with incomplete records, several generations of refurbishment and a service void that nobody has accessed. Producing fluent text about that building is relatively easy. Establishing what is actually there, what remains unknown and how to investigate safely is much harder.
Will AI replace asbestos surveyors?
Our view is that complete replacement is unlikely in the near term. Asbestos surveying combines physical investigation, controlled sampling, building knowledge and professional judgement. AI can assist with parts of the work, but that does not make the whole role autonomous.
Asbestos analysis also contains tasks that technology may automate, alongside practical and quality responsibilities. “AI-resistant” is a more realistic description than a promise of permanent job immunity.
This distinction between a task and a whole job matters. The International Labour Organization’s 2025 research concludes that transformation is more likely than wholesale replacement for most jobs exposed to generative AI. That is useful context, although it is not a forecast specifically for asbestos careers. Read the ILO research.
Seven reasons asbestos work is difficult to automate
1. The evidence has to come from the real building
AI can interpret the information it receives. It cannot make an uninspected area become inspected.
A photograph may show the face of a panel while revealing nothing about the materials behind it. Old drawings may omit later alterations. Even an excellent digital model is limited by the evidence used to create it.
Our assessment is that this need to establish reliable facts on site is one of the strongest barriers to full automation. Faster report writing does not remove the need for investigation.
2. Sampling is a physical task with real consequences
Obtaining useful evidence involves more than selecting a point on a photograph. HSE advises that sampling should be undertaken by a suitably trained person and that suspect materials are confirmed through sample analysis. See HSE’s survey guidance.
For automation, the challenge is the whole physical workflow: reaching the location, recognising the conditions, handling materials appropriately and responding when the situation differs from expectations. A chatbot and a robot capable of safely performing that work are very different technologies.
3. Buildings contain exceptions, alterations and uncertainty
Imagine two identical-looking rooms. One was refurbished years after the other. Their visible finishes match, but their concealed construction may not.
That hypothetical example illustrates why pattern recognition alone is insufficient. A useful system must identify when its assumptions are unreliable, and a competent professional must decide what further evidence is needed.
Survey planning, carrying out the inspection and reporting the findings are all addressed in HSG264, HSE’s asbestos survey guide. The challenge is applying a sound method to the building in front of you.
4. New technology has to fit a regulated quality system
Regulation raises the standard a new method must meet. It does not make technological change impossible.
In Great Britain, HSE strongly recommends UKAS-accredited surveying organisations working to ISO/IEC 17020, while explaining that surveying accreditation is not a universal legal requirement. Commissioned asbestos sample analysis and air measurements have specific ISO/IEC 17025 accreditation requirements. Read HSE’s accreditation guidance.
Licensing is a separate matter, applying to defined licensable asbestos work. It should not be used as a blanket description of every surveyor or analyst. See HSE’s licensing explanation.
The practical implication for AI is straightforward: a convincing demonstration is not enough. An organisation needs evidence that a tool is suitable for its intended use, with appropriate checks, records and responsibility for the result.
5. An asbestos analyst’s role extends beyond a test result
Laboratory analysis and site-based analytical work should not be treated as one identical automation problem. Their tasks and working conditions differ.
For example, four-stage clearance includes checks on site condition, a thorough visual inspection, air monitoring and a final assessment after the enclosure is dismantled. HSE emphasises the importance of the visual inspection. See HSG248, Asbestos: The Analysts’ Guide.
Even if technology improves image interpretation or counting, that does not automatically replace the wider site assessment. Each part of the process needs to be evaluated on its own merits.
6. Someone must challenge an incomplete answer
A polished report can still be based on incomplete evidence. The valuable question is often: “What have we not established?”
Our view is that professionals who can recognise contradictions, explain limitations and challenge pressure to reach a convenient conclusion will remain valuable as software improves. AI may help identify inconsistencies; accepting or resolving them still needs a defensible process.
7. Full automation has to work economically as well as technically
Demonstrating one automated task is different from operating a reliable service across schools, homes, factories and difficult-to-access buildings.
A replacement system would need to justify its equipment, deployment, maintenance, supervision and exception-handling costs. This is an economic argument, not a measured prediction of when robots will become competitive.
It helps explain why assisting a competent person may be a more practical route than attempting to remove that person from the entire workflow.
AI can help write what the survey found. It cannot turn an uninspected void into an inspected one.
Can asbestos scanners replace sampling and laboratory analysis?
Portable sensing technology deserves serious attention. However, “asbestos scanner” describes a range of technologies, not one device with one accuracy figure.
A review commissioned by the NSW Office of the Chief Scientist and Engineer discusses different sensing methods and their limitations. It does not establish a universal performance score for every product and asbestos-containing material. Read the asbestos sensing review.
It is also inaccurate to say that every sensor must physically touch an individual asbestos fibre. For example, the microPHAZIR AS specification describes a non-contact diffuse reflectance measurement method. Read the manufacturer’s specification.
The questions that matter more than “90% accurate”
- Which materials and asbestos types were tested?
- What is the detection limit for that application?
- How does it perform with low or uneven asbestos content?
- What happens when the relevant material is painted, coated or concealed?
- How often does it miss asbestos that is actually present?
- Is it validated for screening, confirmation or another specific purpose?
A headline accuracy percentage does not answer those questions. A negative reading must be interpreted within the method’s validated scope and limitations.
Paint, coverings and low concentrations should therefore be assessed against the particular device and method. Neither “all scanners fail” nor “a quick scan proves it is asbestos-free” is a defensible blanket statement.
For current practice in Great Britain, HSE states that definitive confirmation requires analysis of a representative sample of the suspect material. See HSE’s guidance on analysis.
Where AI could make asbestos professionals more effective
There are useful opportunities to reduce administration and support quality checking. The examples below are potential applications, not claims that every available product can already perform them reliably.
| Potential AI assistance | Professional check |
|---|---|
| Drafting report text from recorded findings | Confirm every statement matches the site evidence. |
| Comparing room lists, plans and sample references | Resolve discrepancies against the original records. |
| Flagging missing fields or inconsistent descriptions | Decide whether the issue is administrative or requires further investigation. |
| Summarising historical reports | Check dates, scope, limitations and changes to the building. |
| Assisting laboratory image review | Validate performance for the specific method and maintain quality controls. |
Better productivity could change staffing requirements even where complete replacement remains difficult. It would be unrealistic to promise that every existing task, role or employer will remain unchanged.
Is asbestos surveying a good career choice in the age of AI?
It is worth considering if you enjoy practical investigation, technical detail and responsibility. Its combination of physical work and judgement provides a credible case for resilience to generative AI.
That is not a guarantee of employment. Consider local demand, working conditions, travel, training support and progression as carefully as automation risk. A career that suits your abilities is a stronger choice than one selected solely because somebody labels it “AI-proof”.
Our advice is to develop the capabilities that make the whole service dependable:
- Building knowledge: understand how construction and alterations affect an investigation.
- Practical competence: build supervised experience and recognise the limits of your competence.
- Evidence discipline: keep records that allow another professional to understand your reasoning.
- Clear communication: explain findings, uncertainty and next steps without giving false reassurance.
- Digital judgement: use approved tools while checking their output and protecting client information.
Our verdict: a strong case for resilience, not immunity
We expect AI to change asbestos work. We see a much weaker case for assuming it can simply replace the surveyor or analyst responsible for the complete service.
The strongest protection is the combination: reliable site evidence, practical skill, technical understanding, quality assurance and judgement when something does not fit.
Nobody can responsibly put a precise number of years on full replacement. For people entering the profession today, the sensible ambition is to become excellent at the work and capable of using new tools critically.
Frequently asked questions
Is asbestos surveying AI-proof?
No occupation can be guaranteed permanent immunity. Asbestos surveying has practical and judgement-based tasks that make full automation difficult, while its administrative tasks are more open to change.
Will AI replace asbestos analysts?
Some analytical tasks may become more automated. That is different from replacing all laboratory and site responsibilities. The effect will depend on the task, validated methods and the controls around their use.
Can AI identify asbestos from a photograph?
A photograph may help flag a suspect material, but it cannot confirm asbestos content or rule it out. HSE’s guidance requires representative sample analysis for definitive confirmation.
Are asbestos surveyors licensed?
Licensing applies to defined licensable asbestos work. It is separate from surveying competence and organisational accreditation. Describing all asbestos surveyors as “licensed surveyors” is misleading in Great Britain.
What makes an asbestos career more resilient?
Strong practical competence, building knowledge, sound records, clear communication and the ability to question unreliable information. Those skills remain useful as the tools change.
Good asbestos decisions start with good evidence
If you need help understanding your building’s asbestos information, survey requirements or management arrangements, speak to ASM about the next step.
Discuss your asbestos requirementsThis article presents ASM’s assessment of automation alongside linked guidance and research. Future employment outcomes are uncertain. Regulatory references concern Great Britain.
