Testing GeneXus Enterprise AI. Testing of applications based on LLMs
07:00 PM - 07:30 PM
Ballroom C
In this talk, we will explore the challenges and different strategies and methods for testing applications based on LLMs (large language models) such as ChatGPT developed by OpenAI, or GeneXus Enterprise AI. As artificial intelligence becomes increasingly integrated into our lives, it is crucial to ensure the quality, efficiency and reliability of applications that use these technologies. This becomes more relevant now that GeneXus also provides access to this type of applications.

The talk will help to better understand the challenges of testing this type of systems, as well as provide some of the applicable testing techniques, and their limitations, existing tools and others that are needed. This is for both functional, automated testing and non-functional testing (performance, security, accessibility, etc.). It is important from now on to reflect on how to face these new challenges, considering that even though they are not already testing this type of applications today, it will not take long for that to happen.
Some topics we will address:
Introduction to LLMs: We will see a brief introduction to LLMs and in particular to ChatGPT and GeneXus Enterprise AI as case studies, their use cases and how they have been integrated into current applications. We will also analyze the quality challenges in implementing these types of software products.
Testing strategies and techniques: We will describe best practices in testing applications based on LLMs, focusing on functional and non-functional tests, as well as the automation of regression tests.
Ethical and privacy aspects: We will reflect on the impact of AI and language models on critical aspects, such as data security, code security, privacy and possible biases.
Conclusions and future perspectives: We will end the talk by analyzing the trends in the field of software testing for applications based on LLMs and other challenges that continue to appear in the world of AI with their corresponding impact on testing.
Note: This text was automatically translated.