In the rapidly evolving world of software engineering, the advent of Artificial Intelligence (AI) has unleashed a wave of transformative change. Recently, Open AI and Google came out with unprecedented AI systems capable of answering questions, writing instructions, providing technical support and way much more.

Concurrently, AI is also shaping the future of software engineering careers, presenting novel opportunities for innovation and growth. For this reason, we aim to unfold how AI is revolutionising the software engineering labor market, by gathering the opinion of two experienced tech executives.

The inevitable rise of AI systems seems very promising, but what are the real impacts that AI is leaving on the software engineering labor market?

To help us unravel the dynamic intersection of AI and the ever-evolving software engineering labor market, we counted on the assistance of two great tech executives to help us delve into the transformative effects of AI on the engineering industry and answer some relevant questions.

But before we dive into the heart of the discussion, allow us to quickly introduce our respondents:

●     Andrew Flegg, CTO at Alfa Financial Software: Andrew has worked on a number of projects within Alfa, providing development, support, management and consultancy to the asset finance industry. As a CTO, he is responsible for the strategy, development and delivery of Alfa’s technology to clients and partners.

Alfa currently boasts a workforce of approximately 500 dedicated professionals and serves a diverse clientele of 30 companies. With a global presence spanning Europe, Asia-Pacific, and the United States, the company has been delivering systems and services to the global asset and automotive finance industry since 1990. In 2022, Alfa’s revenue rose 12% to £93.3 million (equivalent to approximately $115.9 million), according to Investors’ Chronicle.

●     Our second respondent is João Freitas, General Manager & Engineering Site Lead at PagerDuty: João Freitas is currently responsible for the PagerDuty’s Lisbon site and multiple engineering groups. He has a PhD in Computer Science and over 30 articles published in the areas of Speech Technologies, Human-Computer Interaction, Ambient Assisted Living, and Crowd-Sourcing.

PagerDuty operates from six strategically located global offices and boasts a talented workforce of around 1,000 dedicated professionals. The company achieved remarkable financial success, as evidenced by its impressive revenue of $370.8 million for the full fiscal year 2023.

Looking ahead, PagerDuty envisions continued growth at an estimated rate of 20% - 22%, aiming to achieve a revenue milestone of approximately $452.0 million for the full fiscal year 2024, as stated on their official website.

Now that our respondents and their influential companies were presented, five thought-provoking questions addressing the seismic impact of AI in the software engineering labor market were discussed with them.

Here are the key insights that emerged from each one of the questions:

1 - With new and more autonomous and developed AI tools to assist in the software development, what would be the role of the software engineer in this process?

With the advent of new and highly autonomous AI tools that continue to advance and assist in software development, we wonder if the role of the software engineer will remain crucial in the overall process.

In the words of João Freitas, General Manager of PagerDuty, while software engineers will remain essential in the software development process, AI is progressively reducing the costs of IT projects and diminishing the demand for a significant number of human professionals.

Andrew Flegg shares a similar point of view when it comes to the value of the software engineer role: “I’ve had the advantage of coding for a very long time, so the act of actually getting the computer to do what I want isn’t hard. Working out what I want it to do – the design thinking – is the hard bit.”

According to both tech executives, while these tools can automate certain repetitive tasks and enhance efficiency, software engineers bring a unique set of skills and expertise that cannot be replicated by AI alone.

Furthermore, software engineers possess a deep understanding of software architecture, algorithms, and system design, allowing them to make informed decisions and trade-offs during the development process.

In summary, while AI tools are valuable aids, the software engineer role remains indispensable in guiding and orchestrating the software development process, leveraging their expertise and creativity to deliver high-quality and impactful solutions.

2 - When do you think we will see software fully developed/coded by AI?

Predicting the exact timeframe for when we will see software fully developed and coded by AI is challenging. While AI has made significant advancements in automating certain aspects of software development, creating a fully autonomous AI capable of independently and accurately developing complex software systems remains a formidable task. Here is what our experts have to say about it:

Alfa’s CTO, Andrew Flegg does not believe software will ever be fully coded by AI, since some vital tasks can still only be performed by humans, but on the other hand, the executive recognises the efficiency brought by the technology:

“Coding itself has gotten easier through more advanced class libraries (when did you last have to write your own linked list implementation, or manually code an HTTP request?) and tooling such as compile-as-you-type, auto-complete, linting and so on.”

In contrast, João Freitas presents a different opinion regarding the matter: “you are already starting to see parts of programs developed by AI (...) It is not yet always helpful in more complex tasks. However, things are moving fast and we might get there sooner than expected.”

It is important to mention that software development involves not only writing code but also understanding intricate requirements, making design decisions, and addressing real-world complexities. Therefore, a complete replacement of human software engineers seems unlikely, at least in the near future.

Certainly, the collaborative efforts of AI and human developers, leveraging each other's strengths, will continue to shape the future of software development.

3 - Which are the pitfalls of software developed by AI? Who would validate the generated codes? How would it be maintained and reused? How do we protect it from vulnerabilities?

Software developed by AI can present several pitfalls that need to be addressed. One major concern is the lack of interpretability and explainability in AI-generated code, making it challenging to understand the underlying logic and debug potential issues.

According to Andrew Flegg one big pitfall with generative AI based on Large Language Models (LLMs) slurped from the internet, seems to be inadvertent copyright infringement:

“People have equated their suggestions to being “inspired” by Stack Overflow answers. However, if someone has added code to Stack Overflow as an answer, they've intended to share it - to inspire, to educate and to inform. It's explicitly under a Creative Commons licence (CC BY-SA).”

Validating the generated code becomes crucial, and human software engineers would play a vital role in reviewing and verifying its quality, ensuring it meets functional and non-functional requirements.

As stated by João Freitas, “the responsibility/accountability is still on the developer, the same way developers used Google for solutions, now they use ChatGPT (or similar). It doesn't mean that you don't need to review the generated code or understand how to optimize it.”

Furthermore, Alfa’s CTO also warns about copyright infringement by AI systems when reviewing code: “Answers generated by ChatGPT or similar, by parsing code that's available on the internet, may be based on source that is leaked, proprietary, GPL, other OSS or similar. Even reviewing code and deciding it's "safe" from a security or behavioural point of view could still result in unintentional copyright infringement.”

Maintenance and reuse of AI-generated code may pose challenges as well, as updates and modifications might require human intervention to understand and modify the code effectively.

Additionally, protecting AI-generated code from vulnerabilities is crucial, and security measures need to be implemented by both AI systems and human developers to identify and address potential weaknesses.

Once again, collaborative efforts between AI and human developers are necessary to validate, maintain, and safeguard AI-generated code effectively.

4 - Do you think software engineers will eventually be fully replaced by AI?

In line with Andrew’s point of view, it is improbable that software engineers will be fully replaced by AI. They are more likely to work in collaboration as concluded in the previous discussions:

“It’s easy to see AI filling a similar role as an IDE in the future. You can still write a complex system in a notepad, but you’re intentionally hobbling yourself. AI will help with the expression of the idea, but the idea itself – particularly for a complex system – will come from software engineering professionals.”

PagerDuty’s tech executive is on board with Andrew’s view. However, João Freitas reckons that when AI teams up with software engineers, it's gonna unleash a productivity frenzy for companies, ultimately leading to a decreased reliance on manpower.

“GenAI is helping developers to be more efficient. Thus, you need less resources to create the same system. The better it gets, the less people you will need. Some dev tasks are more susceptible to be automated than others.”

But there’s no need to panic. While AI has the potential to automate certain aspects of software development, the role of software engineers extends beyond coding.

Software engineers bring a unique set of skills such as critical thinking, problem-solving, creativity, and domain expertise that cannot be easily replicated by AI systems.

Both executives come to the conclusion that the collaboration between AI and software engineers is more likely to augment and enhance the software development process rather than completely replacing human expertise.

5 - Finally, is the road towards AI inevitable?

The road towards AI can be seen as an inevitable trajectory given the rapid advancements and integration of AI technologies in various domains.

The potential benefits and capabilities offered by AI, such as automation, data analysis, and decision-making support, have captivated industries and societies worldwide.

However, as pointed out by Andrew Flegg, we've still got a long road ahead before we can fully unleash the true potential of AI: “we’re basically at “show me how to do this small thing in language X using library Y.” That means we’re a very long way from “modify this finance system to handle the move from LIBOR to SONIA.” Adding his perspective, the General Manager of PagerDuty states: “we've been on the road to AI since the 50’s."

While the road towards AI seems inevitable, it is crucial to proceed thoughtfully, ensuring responsible and transparent development, while prioritising the ethical use of AI to maximize its positive impact on individuals and communities.