Recent breakthroughs in machine learning have enabled artificial intelligence to reliably and accurately generate text, detailed images, and even high-quality videos. Given that businesses will continue to look for ways to automate jobs and cut down on labor costs, many are now wondering: Will artificial intelligence take your job? Ultimately, the answer depends on the specific job and timeframe that we’re talking about.
In the short-term, jobs involving repetitive, tedious tasks will quickly be completely automated by artificial intelligence. Customer service positions, which simply require the AI model to be trained on a body of information about the given company, have already begun to be replaced by AI-powered customer service chatbots. Even in 2022, before the rise of Chat GPT, research company Gartner predicted 25% of companies would use AI for customer service by 2027. Truck drivers and taxi-like services will eventually no longer be necessary as self-driving technology becomes better and better. Research-heavy jobs such as paralegals, market research positions, and financial analyst positions will likely be automated by artificial intelligence in the future to a heavy degree. All of these jobs have been listed by Business Insider as the most likely to be replaced by AI in the near future.
Surprisingly, software engineering also falls into this category of repetitive, tedious tasks that are at risk of automation through artificial intelligence. Following Cognition Labs’ March release of Devin, an “AI software engineer,” lots of computer science majors had reason for concern about their future job prospects. Devin scored a record breaking 13.86% task accuracy on SWE-bench, a dataset consisting of daily programming issues that a software engineer might encounter while working. In comparison, OpenAI’s GPT-4, which ChatGPT runs on, scored just 1.74% on the same dataset. Although 13.86% accuracy is still too low for widespread commercialization of artificially intelligence software engineers, this number will almost certainly rise in the near future as researchers discover more algorithmic improvements to these models and companies manufacture increasingly powerful GPUs to train them.
What will happen to jobs in the long-term? Unfortunately, the future is incredibly uncertain. One key variable in this question is the development of artificial general intelligence (AGI). As Wikipedia aptly defines it, AGI is “a type of artificial intelligence that can perform as well as or better than humans on a wide range of cognitive tasks.” Such an AGI system would function similarly to a human, only with enhanced cognitive abilities. This would almost certainly dispense with the need for any human labor whatsoever and completely uproot the traditional work-based style of living that has been present for the entire history of humanity. Even jobs requiring physical labor, which have traditionally been thought to be immune to the threat of AI, would be replaced by AGI-equipped robots.
When will AGI be developed, then? In a survey made in 2024 of 1700 AI experts conducted by AIMultiple, the majority predicted that such technology would be developed by 2060. Some predict that this process will be rapidly expedited by so-called “autodidactic systems,” where smarter-than-human AI makes algorithmic improvements to itself to become smarter, which allows it to make even more advanced improvements to become even smarter, and so on and so forth in an infinite feedback loop of acquiring infinite intelligence. This would be a very scary scenario. If this AGI was armed with a powerful processing unit and some way to obtain physical presence in human society, we could very quickly witness the likes of The Terminator and Age of Ultron.
Regardless of whether or not this autodidactic hypothesis turns out to be true, the general consensus is that the development of AGI is unstoppable, and that such an AGI would make human labor useless. This begs the question: How will people continue to make a living in the age of automation? Some have proposed universal basic income (UBI), a government program that would indefinitely provide the citizens of a country with a baseline “wage” for doing anything they please. This would allow people to continue living in a quasi-normal fashion, and the money could come from the profits produced by taxing the corporations that benefit the most from the AI boom.
Even so, it is still difficult not to worry about the tremendous amount of power that the first company to develop AGI would hold. Even if the autodidactic prediction of AGI does not come to pass, this company would have access to immense reservoirs of labor unswayed by ethical considerations. This could be used to enforce a global corporate totalitarian state, with robot police officers and AGI-powered surveillance and flagging tools. Casting dystopian thoughts aside, we must all act together to ensure that safe development of artificial intelligence occurs, so that we don’t accidentally make ourselves unnecessary.