AI: Educating the Educators
The Chemeca conference in Australia had more questions than answers on how AI should best be harnessed to prepare graduates for the future. Although initially discouraging, a group of academics from the University of Queensland believe it shows the way forward
Quick read
- Engage Students in AI Conversations: Educators should create open dialogues with students about their use of AI tools, fostering collaboration and critical thinking to enhance learning and better prepare them for industry demands
- Cultivate Generative AI Literacy: Developing skills in using AI effectively - such as crafting prompts and critically evaluating outputs – is essential for students, ensuring they can integrate AI responsibly into their engineering practices
ARTIFICIAL INTELLIGENCE is a trend that is not going away, and simply telling our students not to use ChatGPT is denying reality. Teaching abstinence doesn’t work; we all need to learn how to teach with AI and learn fast.
At the recent Chemeca conference in Gold Coast, Australia educators, students and industry participants from around Australia and New Zealand met to share best practice in how we harness the fast-evolving AI technologies and tools.
The workshop discussions highlighted that as a chemical engineering profession we are still identifying the best opportunities for use of AI tools in our work and that there is a wide range of adoption of these technologies across and within organisations.
Despite these variations, adaptability and critical thinking skills remain essential graduate outcomes and as educators we need to consider AI as a partner to enhance learning experiences and foster innovation and problem solving.
What chemical engineering academics must do now
Open a dialogue with your students about how they are using AI
Students, academics, and industry are all navigating an environment which is rapidly changing. In this situation, open dialogue is critical. Rather than wait until you are an AI expert (a day which may never come), you need to be having conversations with your students now: how are they using AI in your class? What is it useful for? What problems do they see?
You need to create a safe space for these conversations, which is good educational practice at the best of times but is critically important at this moment. Your students want to be well-prepared for industry, they want their degree to have value. AI therefore provides the opportunity for us to work in partnership with our students as we navigate the questions which concern all of us (see boxout below).
Key questions about AI
Exercise: put these into your favourite AI and see what answers you get! If you haven’t yet used AI, download ChatGPT (or equivalent) and start now by asking it these questions:
Students
- What does AI mean for my career?
- How do I use AI to learn effectively?
- Is using AI cheating?
- Do I need to be an AI expert?
Industry
- How do we harness AI to do what we do better when we’re not AI experts?
- Can we protect our data while
using AI? - How do we avoid getting left behind?
- Who can help us do this?
Educators
- Can I use AI to enhance learning when I’m not an AI expert?
- How do I ensure academic integrity?
- What skills do our graduates need?
Start playing with AI
We need to learn through doing, just like AI, discovering and pushing its potential further.
Find and share good resources
Within any organisation, there are individuals who are following AI advancement with excitement and are keen to share. Sometimes they may need a translator! You can also partner outside chemical engineering, eg to computer science or other fields, to find a partner who can point you to the appropriate resources. For example, The Economy of Algorithms1 newsletter gives an extremely clear overview of where AI should be used, what to keep in mind while using it, and ongoing considerations. For those of you who have an interest in AI, this is your moment; start connecting and sharing with your colleagues.
Beware of AI’s limitations, especially biases and risks
AI’s susceptibility to biases and hallucinations raises ethical concerns in chemical processes, especially regarding safety. The lack of transparency in AI decision-making could compromise safety protocols in hazardous operations. As Decardi-Nelson et al reported in their article in Frontiers in Chemical Engineering on generative AI, collaboration with industry and regulators is essential to establish clear guidelines for ethical and secure AI use.
Where do we go from here?
Whether we are students, academics or practicing engineers, we need to recognise when and how AI tools can be used effectively, and how to design systems which use AI to their advantage. Many have already adopted AI to overcome past challenges that were resource-intensive and time-consuming, with slow feedback on changes.
However, this does not negate the need for core chemical engineering skills: it’s more important than ever that chemical engineers have the fundamental skills to check and verify any work (whether from AI or other sources), and to apply critical thinking to AI outputs. How we educate chemical engineering remains paramount. Building students’ teamwork, leadership, creativity, and communication skills are more important than ever, as AI’s role in engineering grows.
AI offers opportunities for:
Personalised learning
In conversations with colleagues, we’ve noticed a stark contrast between students who effectively use AI tools and those who don’t. One professor shared how students asking vague questions tend to receive overly complex answers, often beyond the course’s scope. However, those who use the AI as a Socratic tutor, where the focus is on asking thoughtful, precise questions, see more meaningful results. They can guide the AI through step-by-step reasoning, engaging in a dialogue much like with a human tutor. This interaction helps develop a better intuition about engineering concepts. As AI evolves into multi-modal environments, students will not only engage with text but also visual data, diagrams, and simulations.
Many students have only explored AI chatbots for writing but haven’t ventured into programming or experimented with AI’s multimodal and code execution environments. Those who do could gain a deeper, more interactive learning experience by instantly testing and validating code. The challenge remains in teaching students how to effectively interact with these AI systems to unlock their full potential.
Engineering drawing
Think about how AI could transform the way we teach process control and design. Imagine a student tackling a complex P&ID assignment late at night, unsure about certain sections. With AI-driven multimodal tools, they could input both text descriptions and incomplete drawings, and the AI could autocomplete the diagram or suggest corrections in real time.
Several groups have already built a novel generative AI methodology for automatically identifying errors in flowsheets and suggesting corrections to the user, ie autocorrecting flowsheets.2 AI’s potential as a Socratic tutor is powerful here too – students could ask follow-up questions, learning the “why” behind each correction, rather than just receiving a fix. Many students don’t realise AI’s capabilities go far beyond writing assistance and can help with technical tasks like engineering drawings. AI’s multimodal functionality – handling both text and visuals – allows it to guide students through intricate design processes.
Data processing
A recurring theme in discussions with fellow educators is AI’s role in handling large datasets. Imagine a graduate student overwhelmed with data from process simulations. Using AI, they could not only analyse the data but ask the system to walk them through the reasoning process step by step, engaging in Socratic dialogue to test their understanding. The code execution environments of advanced AI tools could help them visualise trends in the data while running scripts directly in the interface, refining the model as they go. We emphasise that while these tools speed up analysis, students must also learn to fine-tune models to proprietary datasets and critically evaluate the results AI produces. As future engineers, it’s critical they learn to balance efficiency with ethical considerations, ensuring AI is used responsibly in professional practice.3
Developing “Generative AI literacy” is critical for our students’ futures, so that our graduates can craft effective prompts, critically evaluate AI-generated results, and iteratively refine these outputs to arrive at more reliable, evidence-based conclusions.
The solution is not to go it alone, but to call on the right expertise and build networks of support with those within – and outside – our profession who are at forefront of AI: universities and industry need to collaborate on AI as an engineering work tool, while professional societies such as IChemE can play a leading role in developing strategies to guide use of AI in chemical engineering work.
As the Borg said in Star Trek: “‘Resistance is futile” but perhaps in the context of education, “ignorance is even more futile” – students and academics must embrace AI responsibly to fully harness its potential.
References
1. https://marekkowal.substack.com/p/chatgptchecks
2. LS Balhorn, M Caballero & AM Schweidtmann (2024). Toward autocorrection of chemical process flowsheets using large language models. Computer Aided Chemical Engineering, pp3109–3114. https://doi.org/10.1016/b978-0-443-28824-1.50519-6
3. J Macina et al. (2023). MathDial: A Dialogue Tutoring Dataset with Rich Pedagogical Properties Grounded in Math Reasoning Problems. In Findings of the Association for Computational Linguistics: EMNLP 2023, pp5602–5621, Singapore. Association for Computational Linguistics
Recent Editions
Catch up on the latest news, views and jobs from The Chemical Engineer. Below are the four latest issues. View a wider selection of the archive from within the Magazine section of this site.
