As artificial intelligence continues to develop and grow, a new research paper suggests that it could impact the day-to-day work of nearly 80% of the country’s workforce.
The paper comes from researchers with OpenAI, OpenResearch, and the University of Pennsylvania. It notes that not only will AI begin to impact the work that America’s workforce completes every day, but as the tech develops, it could augment or even replace some labor.
The concept of automating skilled labor is not new by any means, as some have argued in favor and against the idea of using it for things like automated driving, which would replace semi-truck drivers. A report that Bloomberg covered last year said that as many as half-a-million jobs could be replaced by self-driving trucks.
When it comes to AI in other jobs, the paper suggests that nearly 80% of the workforce could see at least 10% of their work changed due to the implementation of large language models.
The paper also predicts that around 20% of workers would see almost half of their work tasks impacted by the same software. Overall, it suggests that 15% of all work tasks in the country could “be completed significantly faster at the same level of quality” if an AI tool was used.
Among the jobs that could be affected by AI include mathematicians, journalists, writers, and authors. The researchers predict that these jobs and others could use software like ChatGPT to cut down on the time it takes to get work done.
While the paper notes it was unable to distinguish “between labor-augmenting or labor-displacing effects” when finding economic impacts of the technology, it concluded that LLM technologies “could have considerable economic, social, and policy implications” if it was introduced to industries in the U.S.
OpenAI is the company behind the ever-growing ChatGPT, which made headlines for how quickly it boomed into success earlier this year.
The language models used in AI, like Google’s LaMDA and the one used for ChatGPT, are known as neural networks. This means that the software can mimic the basic architecture of a brain but in the form of a computer. This allows for substantial amounts of text to be fed to the language model so that it can generate sentences.
The ability to understand the context of words and generate sentences has allowed chatbots like ChatGPT to be used in creating several different forms of content, and eventually, this could be transitioned into the workforce. Users have created essays, screenplays, song lyrics, and even job applications on the platform.
Still, there are several worries that remain present with artificial intelligence and the things that it can do. Most recently, photos that were completely created by AI showed former President Donald Trump being arrested, creating a frenzy online until they were proven fake.
Other concerns revolve around the use of the programs when it comes to school, as ChatGPT and other chatbots have allowed for the creation and completion of homework, adding a new layer to academic integrity in the modern age.
Overall, the authors of the paper, who used AI to help complete their research, shared that “further research is necessary to explore the broader implications” of integrating AI software into the workforce.