The combine harvester is a significant form of job automation, responsible for reducing global malnutrition rates. image source
Over time, jobs have become easier to do thanks to automating tedious tasks. Job automation has allowed us to reduce time in manual labor, and pursue more technical careers. However, with AI on the rise, some might raise this as a concern.
AI can already write articles, draw pictures, create videos, give advice, and so on. According to readwrite.com, during a period from November 1, 2022, to February 14, 2024, writing jobs in the freelance platform Upwork have declined by 33%. Translation jobs have also decreased by 19%, along with customer service at 16%.
Could AI eventually outsource all white-collar jobs and send people back to manual labor? This has been a claim made by some people over time.
How could job automation take place in white-collar industries?
While AI is on the rise, it is not efficient. It can’t make good enough videos to post on YouTube, Instagram reels, etc. What it can do is give a good foundation, while a human editor is needed to make a viral-worthy video.
Articles that AI writes, especially required for SEO purposes aren’t good enough. Keyword density is too high and sometimes the word count demanded by a person will not always be met.
Clients for freelance writers also have software that determines if an article was written by AI. An AI-written article can damage a company’s search engine rankings, as search engines can deem an AI-written article as spam. Unique articles are crucial for making it to the first page of a search engine. This is how freelance writers are currently safe from AI, as long as they can make better content than AI.
Could AI become even smarter?
The only way AI can completely replace freelancing fields, or any white-collar profession in general, is if it’s smarter than a human. There is no agreed date on when this milestone will be met.
Check out these three stages of AI development: ANI, or artificial narrow intelligence, AGI, or artificial general intelligence, and ASI, or artificial superintelligence.
Artificial Narrow Intelligence
This is currently our stage of AI. ANI is designed only to do specific tasks like:
- Being a voice assistant like Siri.
- Weather forecasting.
- Facial recognition like Face ID.
Artificial General Intelligence
This would be a milestone in AI development. AGI possesses human-like intelligence and can perform any intellectual task a human can do. Here are some possible tasks:
- Write articles, and create videos with very good quality.
- It can be installed on cars for complete autonomous driving.
- Solve unknown problems in science or math.
Artificial Superintelligence
ASI is an extremely advanced AI level that greatly exceeds the smartest people in intelligence. It is purely hypothetical. If it exists, it would be expected to have a very high computing speed, and would also improve itself without human intervention, at an exponential rate. Here are some expectations of ASI:
- Solve great mysteries in science, while discovering more concepts that we never knew existed.
- Help develop new technologies that were once deemed impossible.
It will master almost anything you can think of, especially in the digital world.
How advanced would AI need to be to operate in the white-collar world?
The best bet for AI to become better than humans in digital fields is AGI. As mentioned earlier, AGI is achieved when AI reaches human-like intelligence. So imagine a person who just graduated high school with a diploma, has a general level of intelligence, and is ready to explore more in a specific field. That is pretty much what AGI is.
It will be able to write accurate articles according to the person’s request. It can edit videos exactly how you want.
So as long as AGI doesn’t exist, digital workers have the opportunity to get better in their fields and outperform AI. Once AGI is achieved, things can get a little tricky. Before we discuss AGI in the workforce as a form of job automation, let’s talk about when it could possibly exist.
When will AGI become a reality?
There is a big debate about when AGI will become a reality. Some claim that it might never happen, while others claim that the next LLM chatbot could be AGI by the end of the decade.
Many others claim that AGI isn’t currently feasible with modern-day technology, and the only hope is that it’s decades or even centuries away.
According to Techpoint Africa, the first inconvenience to AGI is flaws in AI training. For example, a significant amount of inaccurate data is being fed to AI models. As the dataset increases, it also increases the risk of training data being AI-generated.
Currently, many articles and other online content are AI-generated, and the content is inaccurate and confusing.
Techpoint Africa also claims that an extreme amount of computational power is needed to make AGI possible.
However, the computer scientist and futurist Ray Kurzweil has predicted it could become a possibility by 2029, and the technological singularity by 2045. Kurzweil has made many accurate predictions, but also many failed ones as well. Many people criticize him for predicting AGI by 2029 and claim that it’s just wishful thinking, especially considering the abovementioned inconveniences.
Can AGI be used for job automation?
Assuming AGI becomes a reality, is it guaranteed to replace jobs? What if workers just use it to do their job, instead of the AI completely replacing a worker?
Historically, job automation has boosted productivity and grew industries. Some people may have been temporarily displaced, but positive effects remained in the long term. Here are a few examples:
- Irrigation systems are a type of automation. Instead of workers wasting time watering grass, they can now focus on more complex tasks. Golf courses are one example. Thanks to irrigation and other automated technology, golf courses were able to grow in size, attract more customers, and maintain or even raise employment.
- Combine harvesters are an automated machine that combines numerous tasks to be done simultaneously. This helps farmers accumulate more space to grow more food. Farm automation has significantly reduced global malnutrition.
- Assembly lines help mass-produce goods to be available to the general public.
So for AGI to aid workers instead of replacing them, it needs to be a digital equivalent of traditional job automation.
The first idea, which is already happening, is chatbots. Many companies focused on customer service have chatbots to assist customers. This leads to reduced waiting times. This can assign human customer service workers to focus on more complex tasks.
Another idea could be data entry. AI can help with this while human workers focus on more productive tasks.
Many low-skilled jobs could be at risk of AI replacement. So a solution would be for high school graduates to immediately train in a skilled field so they can jump right into their careers without wasting time on low-skilled work.
Conclusion
To answer the question in the title, as long as AGI doesn’t exist, human workers can use ANI to boost their productivity. This means that we shouldn’t be concerned for now. Concerns should arise once AI becomes smarter, putting highly skilled careers at stake.
Even then, there is no guarantee that AGI would be a bad thing. Advanced AI could be a good tool for education to learn even more skills that will arrive due to the prevalence of AI. Many predict that if AI were to replace anything, several new jobs would be created to compensate for the losses.
Sources
https://readwrite.com/ai-influences-decline-in-freelance-writing-and-service-jobs-study-finds/
https://cloud.google.com/discover/what-is-artificial-general-intelligence
https://techpoint.africa/2024/07/09/artificial-general-intelligence-hallucination/