Testers will inherit the world of AI
There is a well-known phrase in most versions of The Bible: "Blessed are the meek, for they will inherit the earth".
I would like to apply the same logic to the future world dominated by artificial intelligence. AI will indeed transform humanity, and whether the current large-language-models (LLM), or their successors, are capable of human reasoning or fall short, then there is no doubt that AI will have a transformational effect on many industries, especially software.
Today's current flavour-of-the-month in the software industry is 'vibe-coding', where a developer can instruct any one of the popular AI LLM's to write the 'low hanging fruit' of software applications (mainly web sites) in minutes rather than the traditional days, weeks or months of their forefathers. Indeed these 'apps' look amazing, however the 'vibe-coder' will spend hours iterating and testing this code, adding features until it approaches, but never reaches, perfection.
Imagine then that these AI generated apps make it into the hands of professional software testers, whose quality assurance (QA) processes protect the business from calamitous rogue software, and ensure that their Line-of-Business (LoB) software provides a real benefit to the corporation. This rigorous approach will typically send bug reports back to the dev-team who will no-doubt feed these into the AI LLM to generate new versions.
At the same time, other business specialists will also have additional urgent requirements, will feed these into the dev-team who will increasingly hope and pray that their advanced 'prompt-engineering' skills will ensure that the AI LLM will understand the nuances of the world of human business.
At some point during this increasingly complex lifecycle, the AI LLM will 'lose the plot' possibly for technical reasons e.g. token memory limitations, or for financial reasons e.g. AI vendor wants more money to 'keep the lights on'. It may start 'hallucinating' i.e. start generating code which breaks all the previously tested features, or indeed be unable to understand the complexity of the beast it and it's prompt-engineers have created.
If, and only if, a critical LoB app emerges from this 'brave new world', then who is the master now?
It is of course the QA team whose testers are now the gatekeepers for the adoption of the AI generated LoB app by their business. It is they who will ultimately inherit the apps generated by AI and the dev-team, and arguably they who understand the business processes so well that they will be the ones to 'vibe-code' the next versions, or successors, to these AI generated apps.
