- Anthropic report finds only 4% of workers use AI for 75% of their tasks
- Software engineers use AI more than other types of workers
- AI is best described as a tech that augments human capabilites
Even though investments in AI continue to grow, only 4% of occupations use AI for 75% or more of their tasks, highlighting just how far there is to go to fully engrain the tech within our workflows, new reports have claimed.
It might not be heavily relied upon, but plenty of workers have tried AI, with more than a third (36%) of occupations incorporating the technology for at least 25% of their tasks.
The figures come from a paper published by Anthropic’s researchers, who analyze over four million Claude.ai conversations.
Not many workers use AI too much
More specifically, Anthropic’s researchers break down exactly how AI tools are currently being used, revealing that nearly two in five (37%) AI users work in software engineering roles.
Two in five (43%) use cases proved to be for automation, but the researchers observed the remaining three-fifths (57%) of AI requests augmenting human capabilities, or in other words, boosting efficiency. This finding directly challenges earlier concerns that AI might replace human workers, and is just one of a growing number of studies supporting the pro-human argument for AI.
Artificial intelligence usage is also more prevalent in mid- and high-paying jobs, where many knowledge workers find themselves. Both salary extremes, such as manual and physical labor, and highly skilled workers like physicians, used AI less.
A separate Microsoft study added that workers who are more confident with a task are among the most likely to apply critical thinking to GenAI’s output, compared with workers who might be less familiar with the task, confirming AI’s power as a human aid and not a replacer.
However, we could be on the cusp of major change. The researchers concluded: “As AI systems expand beyond text to handle video, speech, and physical actions through robotics, and as AI agents become more capable of carrying out extended tasks autonomously, the nature of human-AI collaboration is poised to transform dramatically.”
