Three AI considerations for tech leaders in Australia
Experts project AI will contribute $280 million in benefits to Australia by 2030. It's no longer just hype; AI is transforming many facets of business and society, including how we work. GitLab research shows that more than three-quarters of organisations globally are already using AI in development or plan to soon. As AI handles more routine tasks in software development, it is already driving improvements for local organisations, including faster software delivery, better developer experiences, and bigger productivity boosts.
But effectively adopting AI isn't just about using new tools. It's about changing the mindset around how teams work, proactively managing risks, and giving people the relevant skills to adapt to new technologies successfully. For Australian tech leaders, the priority is clear: embrace innovation but keep humans front and centre.
Rethink AI risk management
One common concern when implementing an AI tool is the possibility of plagiarism or using copyrighted training data. This year, we may see lawsuits against LLM providers who cannot prove that their models were not trained on copyrighted data or who fail to adequately protect intellectual property.
Companies must thoroughly evaluate providers based on how transparent they are about the data used to train models and their guarantees for data protection. Without considering how AI tools store and protect proprietary corporate, customer, and partner data, organisations may be vulnerable to security risks, fines, customer attrition, and reputational damage.
These risks will serve as a forcing function for AI solution providers to proactively incorporate data protection principles into models from the outset, ensuring ongoing compliance and mitigating the potential for future legal disputes.
Empower developers to upskill
According to a recent Tech Council of Australia (TCA) report, AI could create 200,000 AI-adjacent jobs by 2030. AI enables developers to shift focus from routine tasks to developing higher-level, critical-thinking solutions that humans execute more effectively. This shift frees up time for innovation and allows developers to expand their technical expertise, learn new skills, and advance their careers.
During this transformational period, it will be more critical than ever that tech leadership works closely with its engineering and development teams. Leaders should regularly check in with their teams to understand how these new tools support their work, as well as any concerns and improvement opportunities.
Leave code testing to AI and focus on complex strategies
We'll see this transition begin with the use of AI for code testing; a responsibility that automation is particularly well-suited for. AI will provide end-to-end testing and gain insights into how well a solution meets the end-to-end intended value. Developers are already anticipating AI's impact on their everyday workflows, with 63% of developers responding that AI will significantly change their role over the next five years.
Since most developers spend just 20% of their time writing code, many opportunities exist to apply AI in software development beyond code generation. AI features like merge request summaries, vulnerability explanations, and issue summaries allow developers to spend time on more challenging tasks.
For example, financial platform Airwallex (a GitLab customer) has seen significant improvements after implementing AI into their DevSecOps workflows, including 8x faster deployments and 4x faster feature delivery.
Australian organisations are moving beyond the initial AI hype cycle to focus on real, measurable business results. Tech leaders must adjust their internal structures and processes to fully unlock AI's value. This includes building frameworks for managing potential risks and providing developers with upskilling opportunities that allow them to take on more strategic tasks. Those implementing these changes will see a strong return on their AI investments and position themselves for long-term success.
