The general assumption surrounding AI is that it has the potential to end up taking away an inordinate quantity of jobs from human beings, but is there actually any truth to this sentiment? A team at MIT sought to find an answer to this question, and their research revealed that AI might not be the job killer that so many people fear it might be.
This study was conducted at MIT’s Computer Science and Artificial Intelligence Laboratory, or CSAIL for short, and it refuted a lot of the assertions made so far. For example, Goldman Sachs has estimated that as many as 25% of jobs can be taken over by AI related automation in just a few short years, whereas McKinsey estimates that 50% of all work will be done by AI by the year 2055.
A poll conducted by UPenn, Princeton, and NYU suggested that 80% of jobs will be taken over by ChatGPT, which just goes to show how pervasive this sentiment truly is. In spite of the fact that this is the case, it might not actually be financially viable to have AI do these jobs according to the MIT report.
The research suggests that AI can indeed automate certain tasks, but that doesn't mean that it can replace jobs related to these tasks. For example, an average of 6% of a baker’s time is devoted to quality control, so if a bakery pays 5 bakers $48,000 a year each, it could save $14,000 on an annual basis by having AI do it.
"We find that only 23% of worker compensation “exposed” to AI computer vision would be cost-effective for firms to automate because of the large upfront costs of AI systems.", highlights study.
However, the system itself would cost upwards of $165,000 a year in maintenance and upkeep, which means that just having humans continue doing their jobs would be more financially suitable. This goes to show that just because AI can do a task does not mean that it will be cheaper, and businesses will be looking at costs instead of just blindly replacing humans. It might be more likely that human laborers will incorporate AI into their work flow which will actually boost productivity across the board in the near to distant future.
Read next: How Tech Professionals Can Prepare for the Future of IT
This study was conducted at MIT’s Computer Science and Artificial Intelligence Laboratory, or CSAIL for short, and it refuted a lot of the assertions made so far. For example, Goldman Sachs has estimated that as many as 25% of jobs can be taken over by AI related automation in just a few short years, whereas McKinsey estimates that 50% of all work will be done by AI by the year 2055.
A poll conducted by UPenn, Princeton, and NYU suggested that 80% of jobs will be taken over by ChatGPT, which just goes to show how pervasive this sentiment truly is. In spite of the fact that this is the case, it might not actually be financially viable to have AI do these jobs according to the MIT report.
The research suggests that AI can indeed automate certain tasks, but that doesn't mean that it can replace jobs related to these tasks. For example, an average of 6% of a baker’s time is devoted to quality control, so if a bakery pays 5 bakers $48,000 a year each, it could save $14,000 on an annual basis by having AI do it.
"We find that only 23% of worker compensation “exposed” to AI computer vision would be cost-effective for firms to automate because of the large upfront costs of AI systems.", highlights study.
However, the system itself would cost upwards of $165,000 a year in maintenance and upkeep, which means that just having humans continue doing their jobs would be more financially suitable. This goes to show that just because AI can do a task does not mean that it will be cheaper, and businesses will be looking at costs instead of just blindly replacing humans. It might be more likely that human laborers will incorporate AI into their work flow which will actually boost productivity across the board in the near to distant future.
- Also read: AI Might Surpass Humans in All Tasks by 2047
Read next: How Tech Professionals Can Prepare for the Future of IT