‘Ethical issues should always be in focus’: How to use AI in engineering

The AI roll-out is already well underway, offering engineers a multitude of ways to optimise processes, boost productivity and lower costs. But in such a fast-moving and complex field, it can be hard to know where to start.

An upcoming IMechE course, AI for Engineers, will tackle that challenge. Running in London from 14-15 April, it is designed to provide an intuitive understanding of how AI works.

The course is led by consultant and AI specialist Ali Parandeh. Here, he gives five key tips to help you start using it effectively. 

Tackle adoption blockers

A lot of organisations are currently looking to adopt AI because they've seen competitors start using it, benefiting from improved profitability and efficiencies in their systems.

While AI promises to transform engineering workflows, adoption is often blocked by a lack of quality data, skilled talent or a clear business case. Engineers should first focus on improving data collection and preparation strategies, upskilling in AI fundamentals and identifying AI use cases with measurable returns on investment.

If you can overcome these adoption blockers, you will have more success trying to utilise AI technologies in your day-to-day workflow. 

Get to grips with the fundamentals

The landscape is evolving and expanding massively. There are up to 350 papers being released every day in this field, but even as new models and applications emerge, the core principles remain the same – data preprocessing, model training and evaluation. Engineers who understand these fundamentals will be better positioned to apply AI effectively across different industries.

There are different types of AI and they each have their own use cases. If you understand each type, it allows you to have a framework to brainstorm ideas.

Make the most of accessible tools

You don’t have to be a programmer to use AI. Platforms like Microsoft Azure Machine Learning Studio and Azure AI Foundry let you build AI models and chatbots without writing code. You can upload your data to these tools and follow their guides, as long as you understand the underlying fundamentals – the concepts of what makes a good model, how to evaluate it, whether it's a good fit for your use case and its limitations. These tools help engineers get started quickly without needing deep technical knowledge.

As we go forward, more engineers will pick up skills for building AI tools, so essentially they could become data scientists in future.

Train up on Python

While no-code solutions are useful, Python remains the go-to language for AI development. If you have spent some time learning how to write MATLAB code, learning Python isn’t difficult. Engineers who invest time in learning Python gain access to a vast ecosystem of AI tools and models, from TensorFlow and PyTorch to Scikit-learn, allowing for more customisation and control over AI solutions.

Don’t ignore ethical issues

With any kind of AI project, especially when it comes to scoping, the ethical issues and the ethical dynamics should always be in focus. AI has the power to drive efficiency, but it also introduces ethical challenges such as bias, fairness and transparency. As a result, engineers should integrate ethical considerations into their projects from the start by following ethical AI frameworks.

IMechE’s AI for Engineers course next runs in London from 14-15 April. Find out more and book on the IMechE training page.

Want the best engineering stories delivered straight to your inbox? The Professional Engineering newsletter gives you vital updates on the most cutting-edge engineering and exciting new job opportunities. To sign up, click here.

Content published by Professional Engineering does not necessarily represent the views of the Institution of Mechanical Engineers.