Cundall Codes Issue 6

This issue: What does Chat GPT mean for engineering consultancies

The release of Chat GPT and other rival language models has sparked waves of excitement and hype, with one tech commentator Yuval Noah Harari suggesting we have summoned an “Alien intelligence”.

But do the practical benefits of this technology meet the hype?

As a language model that understands coding, Chat GPT is adept in interpreting an engineer’s requirements. Whether that’s Excel formulas, Visual Basic for Application (VBA) code or Python code, it can advise on the correct syntax for their engineering problem. It can also generate technical content in a more concise and digestible format. For example, helping write key statements in a report to ensure that the main takeaways are clear for the reader.

Another benefit of the explosion of Chat GPT that is often overlooked is its impact on a company’s overall approach and excitement for digital innovation.

The hugely accessible nature of Chat GPT means anyone capable of using a web browser now has access to AI chatbots that can help them write better reports or start their coding journey.

Those that were previously intimidated by coding can now have large swathes of code written for them. On top of that, the chatbot can explain the reasoning behind each line of code at a pace that’s comfortable to the user.

For those already knowledgeable about coding, Chat GPT allows them to improve the structure of their code, write code quicker or generate higher quality code for them to use. As their expertise grows, more ambitious projects such as API integrations into other tools can be considered.

At Cundall, we recently enjoyed the launch of our internal annual coding competition Creative Fest. Three entries into the competition involved open AI integrations into existing processes. From lighting design to embodied carbon calculations, it’s exciting to see people across all areas of the business experimenting with tools that assist in their day-to-day engineering.

However, when it comes to strictly engineering problems, Chat GPT is a mixed bag. Engineering commentary given by Chat GPT is known to be inconsistent, fabricating references to journals or textbooks or being simply incorrect. A further challenge lies in the explicitly public nature of ChatGPT, as all content generated through it becomes part of its continual learning data set. Engineering consultants should therefore never ask a question that involves project specific material, as that is a fundamental breach of client confidentiality. This makes it difficult to ask Chat GPT meaningful engineering questions to generate an effective solution to their problem.

To address this, Cundall has already set out a comprehensive internal document that outlines best practice for using ChatGPT. It also includes strict enforcement of not using project specific material as an input.

With more AI rivals entering the scene, the availability and accessibility of this technology is too powerful to be ignored. It’s allowing staff from all technical backgrounds to participate in a tech revolution. Typical processes and workflows common across engineering consultancies are being positively disrupted at a pace never seen before.

Consultancies not fostering a strong digital culture and encouraging their staff to embrace the AI boom, whilst maintaining best practice, are going to be left behind.

If you would like to get in touch about this or any other topics of digital engineering, please contact csdigital@cundall.com.