What can ChatGPT tell us about dimensional metrology?

The robots are here for our jobs. Or so we’ve been told for what seems like my entire life now. From Rosie, the Jetson’s maid, to the Roombas vacuuming our living rooms, to the order kiosks at fast food restaurants; the promise of automaton assistants has always seemed to be right around the corner. But should we be fearful, or hopeful (or both)?

I recently had the opportunity to access ChatGPT, an AI language model that generates the most natural-sounding text of any AI writing I’ve read to date. Having previously played with the DALL-E AI image generator, I knew to keep my expectations low but was excited to see what AI knew (or at least thought it knew) about the field of metrology. So, I asked it a series of questions on the topic. Here’s my conversation with ChatGPT and some commentary added by yours truly.

How ChatGPT defines metrology

Sean: What is metrology?

ChatGPT: Metrology is the science of measurement and involves the precise determination of the size or shape of objects or the degree of accuracy in their manufacture. Metrology is used in a wide range of industries, from engineering and manufacturing, to biomedicine and nanotechnology. Metrology-related fields include dimensional, electrical, optical, thermophysical and time measurements.

So far, so good. The answer is a bit basic, but so was the question. Fortunately, ChatGPT references previous questions and responses when formulating subsequent replies in a conversation, so we add context and detail with further questions. A more complete answer would mention metrology’s use in improving part design and process quality and sustainability.

Sean: Who defines the standards followed by metrologists?

ChatGPT: The standards followed by metrologists are typically defined by international organizations and forums such as the International Organization of Standardization (ISO), the International System of Units (SI), and the International Electrotechnical Commission (IEC). In some cases, they are also established by national standards organizations such as the American Society for Testing and Materials (ASTM).

Also good, but these things are easily Googled, so let’s keep pushing further.

Sean: What does a dimensional metrologist need to be effective at his or her job?

ChatGPT: For a dimensional metrologist to be effective in their job, they need to have an understanding of the principles of metrology, a solid foundation in mathematics, a keen eye for detail and accuracy, and a basic understanding of statistical analysis. They must also be able to use specialized metrology equipment like micrometers, calipers, and microscopes. Additionally, they should possess strong communication, problem-solving, and interpersonal skills to interact effectively with other professionals in the workplace, as well as with customers and clients.

Now we’re talking! All the aforementioned tools and skills are useful to metrologists. Let’s focus a little bit more on the equipment used by metrologists.

Sean: What factors should a dimensional metrologist consider when selecting measurement tools?

ChatGPT: When selecting measurement tools, a dimensional metrologist should consider the accuracy and precision needed for the task at hand, the size and shape of the object to be measured, the number and type of measurements that need to be taken, the complexity of the measurements, and the budget for the tools. They should also consider the compatibility of the tools with their existing metrology infrastructure.

Wow, seriously impressive but let’s keep pushing.

Sean: How do accuracy, resolution, speed and other factors influence metrology equipment selection? What tools are commonly used in different metrology for manufacturing industries?

ChatGPT: Accuracy, resolution, speed, and other factors all influence metrology equipment selection. Generally, accuracy and resolution are of the utmost importance, while speed and cost become more important when a large number of measurements need to be taken in a short time-frame. Commonly used tools in different metrology for manufacturing industries include micrometers, calipers, lasers, and coordinate measuring machines (CMMs). Optical and tactile (contact) probes are also used to measure a wide variety of features, depending on the application.

I think we’re approaching some kind of knowledge limit here, as ChatGPT’s starting to repeat itself a bit, but it’s apparent that the AI has a good head on its figurative shoulders when it comes to informed decision-making. Can we keep pushing this topic or have we hit a dead end?

Sean: No, tell me specifically what tools a dimensional metrologist working in aerospace, or automotive, or energy industries would use. How do large or small part or feature sizes create challenges for reliable measurement?

ChatGPT: A dimensional metrologist working in the aerospace, automotive, or energy industries may use tools such as micrometers, calipers, optical and tactile (contact) probes, vision systems, coordinate measuring machines (CMMs), lasers, and other specialized equipment specific to the industry. Large or small part or feature sizes can create challenges for reliable measurement as the right tool must be selected in order to ensure accuracy and resolution. For example, vision systems are commonly used to quickly measure very small parts or features, whereas micrometers or calipers may be necessary to measure larger parts with greater accuracy.

Yeah, good logic and a bit more detail on specific tools, but I’m either not asking the right questions or ChatGPT’s knowledge on the topic only goes so deep. It seems that to gather more information on the specific challenges of inspecting different size features, more specific.

Sean: How can metrology results be used to inform manufacturers on corrective process or tooling changes?


ChatGPT: Metrology results can be used to inform manufacturers on corrective process or tooling changes by highlighting areas of variance from the required specifications. By analyzing the results, manufacturers can identify patterns or trends that can be used to determine the best corrective action. Additionally, the results can be used to validate the effectiveness of corrective actions and verify that components meet the required specifications.



So, we’ve struck upon the sometimes cyclic/iterative nature of manufacturing, hinting at back-projection of deviation plots onto the inspected part, as well as trend analysis. Since ChatGPT seems to have decent decision-making logic, I wondered if it could expand on tool selection methods.

Sean: When selecting new measurement tools, what methodology can a metrologist use to validate the tool is appropriate for his or her application?

ChatGPT: When selecting new measurement tools, a metrologist can use a methodology such as Design of Experiments (DOE) to validate that the tool will work for the given application. DOE involves running controlled and systematic experiments with different parameters to determine which combination of parameters generates the desired result. This allows the metrologist to accurately assess the suitability of any given tool for a given application.

Not bad. If I knew little to nothing about a subject but could formulate a few questions for ChatGPT, it could certainly steer me in the right direction or at least provide some context on what to research for better understanding of the topic.

What can we conclude from this experiment?

Of course, ChatGPT is just a model and there are plenty of limitations on the data used to train it that result in limited detail or in some cases, wrong or unreliable information. The OpenAI website goes into more detail about this and is offering a prize for people who submit examples of ChatGPT’s bad behavior. Unfortunately, I wasn’t really able to find any fault with its replies on this subject.

As with all new technologies, early adoption comes with its share of risk that must be considered on balance with the potential that technology brings. Understanding the limitations of the technology and managing their impact on your process is crucial for their successful use.

And maybe it’s important to know that while computers may be faster than us at many things, they’re not usually smarter than us. Having a human in the loop is critical to mitigation of errors in the implementation of that technology before it can be trusted to its own devices. A tool is only as useful as the operator’s ability to use it reliably.

I think we can agree that while I probably won’t be handing over decision-making abilities to ChatGPT when selecting a new measurement tool anytime soon, there are other avenues where my work-life can benefit from AI. Blog post creation, for one.

If you’re interested in learning more about how AI is transforming the manufacturing industry, check out our blog on AI-based automation or find out how to automate your production with Oqton's Manufacturing OS.

And now for the actual blog post written by ChatGPT summarizing the above conversation...

---

ChatGPT's blog about metrology

The need for high precision and accuracy in industrial processes such as manufacturing, aerospace, automotive, and energy production necessitates the appropriate selection of measurement tools. To ensure that the right tool is chosen, metrologists can use a methodology such as Design of Experiments (DOE). This involves running controlled and systematic experiments to identify which combination of parameters produces the desired result.

Thus, DOE allows metrologists to assess the suitability of any given tool for a given application, especially when precision, resolution, and speed are of the utmost importance.The right tool must be selected to ensure accuracy and resolution. For example, vision systems are commonly used to quickly measure very small parts or features, whereas micrometers or calipers may be necessary to measure larger parts with greater accuracy.

Moreover, metrology results can also be used to inform manufacturers on corrective process or tooling changes by highlighting areas of variance from the required specifications.

By analyzing the results, manufacturers can identify patterns or trends that can be used to determine the best corrective action. Additionally, the results can be used to validate the effectiveness of corrective actions and verify that components meet the required specifications.

In conclusion, using the DOE methodology is essential for the appropriate selection of a measurement tool when precision, resolution, and speed are of the utmost importance. This allows metrologists to accurately assess the suitability of any given tool for a given application. Furthermore, metrology results can be used to inform manufacturers on corrective process or tooling changes and validate the effectiveness of corrective actions.