A venture-capital-funded gold rush in software development has seen nearly $1 billion (£760 million) invested in AI coding assistants, but skilled human babysitters are starting to tire of their dim-witted robot sidekicks.
“I stopped using AI because I was getting stupid,” one senior developer explained to robotics expert and AI critic Philip Piecniewski last week.
Researchers from Microsoft and the University of Edinburgh found that reviewing and double-checking AI output took time, distracting from “other productive tasks, like writing code or running tests.” As a result, the researchers reported, “increased cognitive load, frustration, and time spent on the tasks GenAI is intended to support.”
In some cases, “using GenAI support can actually lead to reduced productivity.”
Moreover, it takes a lot of experience to spot the errors that AI creates, Piecniewski said: Novice developers who outsource their work to AI will lose the ability to become more experienced themselves.
Think about it. That's one way Forster's machine society would degrade itself. Google researchers also warn of another way the machines could grind to a halt: Low-quality AI-generated code that quickly becomes difficult to understand and maintain, accumulating what the jargon calls “technical debt.”
Perhaps we should think of E.M. Forster's story not as fiction, but as a practical manual. Ironically, our thought leaders have gone in the opposite direction, retreating from reality to fiction. Today's AI policy experts prefer to speculate about hypothetical AIs and invent fictional problems.
They convinced Rishi Sunak to host a global summit to address the existential risk of killer AI, something that nobody has yet or will ever invent. The reality is, he barely intervened. This is really a huge failure of intellectual leadership.
One insight from aviation safety guru Professor Wiener seems especially relevant. Professor Wiener devised a list of 31 frivolous yet astute safety rules. The first 16 are left blank. Professor Wiener's 20th rule was, “Are you off guard? Don't worry.”
