It’s a question that’s currently ricocheting throughout the IT industry: “Will AI replace software engineers?” While we’re not into crystal balls and mystic visions, we can say from our observations as tech recruitment specialists that the answer to this question is a resounding ‘No’.
AI appears impressive at some general coding tasks and data processing – at first glance. But there are core aspects of software development where human ingenuity and emotional intelligence just can’t be beat, particularly in tech lead jobs. Let’s explore the key reasons why AI is no match for human-centric skills.
The Pulse Check on AI Usage in Australian Tech
The current discourse around AI makes it easy to believe that it has software developers shaking in their boots. But it’s really not the case.
Stack Overflow’s 2023 Developer Survey shows 66% of Australian software engineers were either using or planning to use AI tools in their work over the course of the year.
They found, however, that developers in hardware, backend systems or applications were less likely to be currently using AI, and likely won’t in the future. This is because the work they do is highly complex, so it’s beyond the AI’s domain. Data scientists, front-end, full stack and cloud infrastructure developers were more likely to be currently using AI tools.
For Australian software developers, the biggest benefit of using AI tools is increasing productivity (33%), followed by greater efficiency (25%), and speeding up their learning (24%). Developers just starting their coding journey said speeding up their learning was the biggest benefit (42%).
Aussie developers are using AI to:
- Write code – 90%
- Debug and get help – 51%
- Document code – 41%
- Learn about a codebase – 30%
- Test code – 30%
While it’s clear developers are enjoying AI’s ability to speed up code writing using tools such as GitHub Copilot, Tabnine and OpenAI Codex, trust remains an issue. They’re split when it comes to AI output – 38% say they trust it and 32% say they don’t. In addition, they don’t feel it will improve coding accuracy.
Where Humans Excel
There are many areas where AI can’t compete with human software developers.
Critical thinking
Critical thinking is an essential part of software development. It helps you to analyse problems, challenge assumptions, evaluate solutions and make informed decisions, all resulting in better-designed software.
A core trait of critical thinkers is being able to see a connection between multiple concepts. For instance, a software development team needs to think critically about balancing feature development with resource constraints. The engineering team wants to implement cutting-edge features to attract users, but the project manager is focused on staying within budget and meeting deadlines. By evaluating the project requirements, timeline and available resources, the team can prioritise features effectively, ensuring both innovation and practicality in the final product. This is something AI simply cannot do.
Software engineers must employ their critical thinking skills to gate-keep AI output. Humans are still required to analyse its logic, verify its accuracy and evaluate its performance against expected outcomes. Additionally, humans can identify potential biases, edge cases and limitations to ensure the code meets quality standards. Most importantly, humans are essential for ensuring an AI’s output aligns with project requirements and contextualising it within broader business objectives. In contrast, AI relies heavily on big data patterns without any contextual adaptation.
Attention to detail
One of the biggest problems with AI output is the high rate of false positives it throws out. Ubisoft’s Commit Assistant is one example. Somewhat ironically, it’s an AI tool that identifies bugs in code. While it catches 6 out of 10 software problems correctly, it has a 30% false-positive rate.
This means software developers must be able to not only vet each flagged piece of code but understand the logic behind how AI came to the false positive. They’re skills that have the potential to only increase the need for software developers, ensuring the answer to the question ‘Will AI replace software engineers?’ remains a firm negative.
Complex reasoning and collaborative problem solving
Another difficulty with AI is its inability to deeply reason about complex real-world problems – something that top software engineers excel at, whether it’s optimising code performance, designing scalable architectures or troubleshooting system failures. Can AI do all that? Not so much.
Solving complex technical problems requires collaboration. Developers are just one part of a multidisciplinary team that brings together diverse human perspectives, expertise and communication styles to brainstorm solutions.
Creativity
An inherent problem with AI is its lack of creativity. But the best programmers know how to blend technical knowledge with creative flair. Consider the creativity required for building an AR or VR interface, for example.
GitHub’s CEO recently said their AI-powered code completion tool, Copilot will write 80% of code sooner than later. He also highlighted that it doesn’t mean the developer will be replaced – it just means they’ll have more time to focus on the 20% they’re writing.
But as many professional developers know, that remaining 20% is tough. It requires an immense amount of grind combined with a bucket load of creativity to produce; skills AI just doesn’t have.
Emotional intelligence
The emotional intelligence of human software developers will always be valued, particularly when it comes to user experience. Your deep understanding of human behaviour and psychology allows for an intuitive and engaging design, producing interfaces that resonate with users and ensure the ultimate end game – they buy your product/service. No machine learning model can compete with that.
Ethical decision-making
From data privacy to algorithmic bias, developers must navigate a complex landscape of ethical dilemmas. While AI can assist in identifying potential ethical issues, it lacks the moral reasoning and judgment necessary to make informed decisions.
As a human developer, you don’t. You can weigh the potential impacts of your work on individuals and society, alongside what’s happening in the wider world, ensuring your software is developed responsibly and with integrity.
Future-Proofing Your Career
Our advice is not to fear the march of AI but embrace the possibilities it presents for your career. AI is automating a sizeable chunk of coding, but it’s also creating a whole new cache of tech and tech lead jobs. Examples include AI software architects, AI DevOps engineers, AI security specialists and AI trainers.
Whether you’re building AI-focused skills or working on your human-centric ones, our specialists here at Exclaim IT can support you. Please contact one of our team to find out more.
