How Much Do AI Models Automate Jobs?

The notion that AI models are poised to usurp human workers has been a hot topic of debate in recent years - but a new study from Google Research throws some cold water on that idea. It turns out that, for the most part, AI isn't replacing humans just yet. The study, which dug into 15 million anonymized AI interactions, found that most tasks in most jobs are still pretty much unaffected by AI automation. What's happening instead is that workers are using AI in a more collaborative way - it's assisting with specific tasks, rather than taking over entirely.

The researchers used an automated classifier to sort through work-based AI interactions and categorize them by occupation. What they found was that while AI is indeed being used in a bunch of different jobs, its use is actually pretty "shallow and overwhelmingly collaborative in nature." It seems that certain white-collar jobs - think financial analysts and software developers - are more likely to leverage AI for job-related tasks, whereas salespeople and transportation workers are less likely to use it.

The study's findings paint a picture of an AI impact that's not quite as seismic as we might have thought. Workers are still shouldering the bulk of the responsibility for most tasks in most jobs. In fact, the researchers found that for a whopping 29% of occupations, not a single relevant work task had been automated using AI - and for another 30% of occupations, less than one-quarter of tracked tasks had seen significant AI usage. This suggests that while AI might be super useful for certain tasks, it's not replacing human workers en masse just yet.

So what's the takeaway from all this? The study's findings offer a more nuanced view of AI's impact on the workforce - and suggest that workers and businesses should focus on using AI as a tool to augment human capabilities, rather than trying to replace them entirely. As AI continues to grow in use, it'll be crucial to keep an eye on its impact on the workforce - and make sure workers have the skills and training they need to work effectively with AI systems.

Source: Ars Technica