This new field is very interesting and almost everyone I know uses it for fun and interest. But now we’re moving to practical applications, focusing on how we can use this technology to transform the way we work and better serve our customers.
“Artificial intelligence made headlines thanks to ChatGPT’s text generation capabilities. But creating software code update content here could also save engineers time.”
There are many applications of AI for knowledge workers. ChatGPT includes 175 billion parameters, making it one of the largest and most powerful models for AI processing available. Also, in January, just two months after the service launched, monthly active users reached his 100 million. fastest growing According to UBS, it will go down in consumer application history.
But what are the tangible benefits of this technology in financial services? How are we driving its use within ANZ? There are many ways to improve , reliability, and performance.
Many people use ChatGPT to generate English text, but it is also very powerful for generating software code snippets. Software is written in various programming languages and, like natural languages, follows a set of grammatical rules.
And there is a huge amount of example code taken from the public domain that trained ChatGPT models to recognize and generate software just like written words.
These examples are peer programmers acting like “over-the-shoulder” assistants, monitoring the writing of code and providing real-time suggestions and feedback on the code being written. This allows engineers to lay out code and build robust outlines more quickly, maximizing common techniques to identify bugs early. This means engineers can spend less time on repetitive coding tasks and more time solving complex problems.
digest complex information
ChatGPT’s conversational AI also helps software engineers understand better. This tool provides answers to common questions and helps engineers find and understand complex technical information. ChatGPT can summarize large amounts of technical information into more understandable chunks, and it can even take engineers’ code and create easy-to-read documentation.
We duly licensed one team at ANZ to research generative AI for software engineering to improve testing. Writing test cases (usually also written in code) is a rather tedious but important task.
Generated AI is really powerful Read the code and suggest a complete test case to make it run properly. I think this will dramatically improve unit testing and make a big difference when it comes to identifying bugs and errors early in the development process.
We are also working to improve how code is validated against strict standards and rules for software code.we are comprehensive ANZ policy However, when completing a code review, a human needs to read and understand the policy and recall key elements.
We see great potential in using generative AI to complete that interpretation. Read policies, read recently added code, and advise where non-compliant areas lie. Of course, it is now tightly controlled.
efficient engineer
We are building a position as we explore this. We primarily see this as an extension tool rather than the final arbiter. All of our code now relies on senior engineers to approve it for completeness and compliance before it gets into the main repository. While that responsibility remains the same, we believe that the use of AI technology can make engineers even more efficient.
Every new technology presents challenges and opportunities, and this also applies to the use of generative AI in software engineering. There are concerns about the reliability, reproducibility, and risks of these tools. Checks and balances will be needed for some time yet. We are not yet ready to use this outside of centrally managed, executive-approved experiments.
However, ANZ has a long history of adopting technology early in the implementation lifecycle. We launched his first website in 1996. Internet banking in 1999. It introduced mobile banking in 2008 and became the first bank to release Apple Pay in 2016.
We have over a third of our application estate in the cloud, run critical workloads in massive container architectures, and employ the latest in data technology. ANZ has been an early adopter of significant new technology and will continue to do so. Generative AI is the next big wave.
But we do it with rigor and discipline.
Tim Hogarth is ANZ’s Chief Technology Officer.
