I was doing the writing section of an IELTS (International English Language Testing System) practice test. Now,
this something A.I. is good at.
Task 1 was "Summarize the table which shows sales of a small restaurant in a downtown business district. Write at least 150 words.
Here's my answer:
A look at this restaurant's sales for the week of October 7-13 reveals that people consistently spend more money on dinner than lunch. The highest revenue for dinner came in on Friday, and the largest revenue day for lunch was Wednesday.
Weekend days brought in less business than weekdays for both lunch and dinner, quite possibly because less people are in the business district on weekends. The highest revenue for lunch was on Wednesday, perhaps because more people need a break from work in the middle of the week, and the highest dinner revenue was on Friday, another day when people want to let loose as a result of working life.
Depending on how much it costs to open the restaurant, it might be a good idea to close the restaurant on weekends. It would also be useful to know how much the lunch and dinner meals cost and how many patrons visited on each day. If the additional dinner revenue comes from drink sales, it may be worth promoting lunchtime drinking, as long as it is ethical in the restaurant's context.
There's actually not much to say about the table, which had 14 data points in it. I was really struggling to hit 150 words. ChatGPT did beat me here:
This table presents the sales data for a restaurant during the week of October 7-13, breaking down the revenue for lunch and dinner on each day of the week. The lunch sales ranged from a low of $1,550 on Sunday to a high of $2,595 on Wednesday. The average lunch sales for the week seemed fairly consistent, staying within a $450 range, with a noticeable dip on Sunday.
Dinner sales were consistently higher than lunch sales, with the lowest dinner total of $3,445 on Wednesday and the highest at $4,350 on Friday. Overall, dinner sales were significantly more stable, with only minor fluctuations throughout the week. Friday was the peak day for dinner sales, and Saturday saw a surge in dinner sales as well, reaching $2,900, which was quite a bit higher than other weekdays.
In terms of overall trends, both lunch and dinner saw a drop in sales over the weekend, particularly Sunday, which showed the lowest numbers across both meals.
The second task to respond to this prompt with 250+ words: "As the world becomes more technologically advanced, computers are replacing people at more and more jobs.
What are some job positions that may be lost because of computers? What are some problems that may result from this situation?"
I probably could have written 5,000 words on this. Here's my response:
As technology develops, computers do replace people at jobs, but not necessarily because they are better at those jobs. The key to job elimination is not doing a better job than people, but rather, convincing investors that they can get a higher return on investment by eliminating people's jobs.
For example, a human customer service representative may do a better job at actually helping a customer than a simple branching script running on an automated phone system or an A.I. assistant. However, the metrics that businesses will often track are: 1) How long the company spends answering calls 2) how many calls answered per day 3) customer satisfaction survey results and 4) how much it cost the company.
The time spent on calls and calls answered per day favor automated systems because customers will often hang up or give up once they realize the system is automated and they won't get the help they want. Similarly, many automated phone systems will simply end the call if they've reached a dead end in their script. As for customer satisfaction surveys, those are often presented at the end of the call. Many customers do not have the luxury to spend extra time on them, and of course, those that do not make it to the end of the call will make it. An automated system will nearly always be cheaper in labor costs than human employees. That is, after all, their central pitch.
All of this amounts to automated systems scoring far higher than human customer service representatives in commonly measured metrics. This looks great in a report, and the effect on company profits will not appear in the short term, as reputation takes a longer time to react to drops in quality than it takes for changes in costs and metrics to appear, which is almost immediately. Investors generally do not operate in individual company investments on a long-term outlook, so it is fine for them if the poor service catches up with the company long after the investor has cashed out.
Similarly, illustration work can and is being replaced by A.I. image generation. In ad design and journalism, graphics are needed to keep the reader's interest among groups of words. Ideally, an illustration will fit its context and make the piece stronger just as much as the words do. Sometimes, computer clip art or stock images are used instead of hiring an illustrator. However, these have become easily identified by typical reader and are either tuned out or harm the reputation of the piece, making it seem "cheap."
A.I. image generation, however, can produce images that mimic the real art that then are trained on (often against the wishes of the creators of that art), that may seem "off" but are not quite the same as any other image out there, unlike stock images or clip art. For businesses making publications or ads, it is often free, thanks to speculative investors subsidizing the cost. Human illustrators charging anything at all can't beat that price, and so the ads and publications will have graphics that will be qualitatively but pervasively worse, lacking intentionality. Their product will feel "garbagey", but again the reputation cost will not land before the upside in cost savings will.
There are some jobs in which computer automation will produce better results than humans working at those same jobs, like searching for proteins with particular properties or precision laser cutting. Of late, however, computer automation is replacing human jobs whether or not it does better work, as long as it can product short-term cost savings.
I realize I didn't talk about the problems explicitly; I feel they're strongly implicit, but maybe I'd get dinged for that. Here's ChatGPT's answer, which does check all the boxes, and of course, assumes that A.I. does do a good job:
As technology continues to advance, many jobs are being replaced or heavily altered by computers and automation. This trend can be seen across various industries, where machines and artificial intelligence (AI) are increasingly performing tasks that were once carried out by humans. Some job positions that may be lost because of computers include roles in manufacturing, customer service, and transportation.
In manufacturing, robots and automated systems are already replacing workers on assembly lines. Machines are faster, more accurate, and do not require breaks, making them more efficient than human workers. For example, car manufacturers use robots for tasks such as welding, painting, and assembly, significantly reducing the number of people required on the production line. This has led to a decrease in factory jobs, especially those that involve repetitive manual labor.
Customer service jobs are also being impacted by computers, particularly through the use of AI chatbots and automated phone systems. These technologies can handle customer inquiries, process orders, and resolve issues without human intervention. For instance, many companies now use AI-powered chatbots on their websites to answer questions, schedule appointments, and solve basic problems. This reduces the need for call center employees and customer service representatives.
The transportation sector is also facing changes, with self-driving vehicles and drones expected to replace jobs in trucking, delivery, and even taxis. Companies like Tesla and Waymo are developing autonomous vehicles that could eventually replace human drivers. This poses a significant threat to truck drivers, taxi drivers, and delivery workers, who may see their jobs eliminated as self-driving technology becomes more reliable and widespread.
The loss of these jobs can lead to several problems. First, there is the potential for significant unemployment, particularly among low-skilled workers who are most vulnerable to automation. This could lead to increased poverty and income inequality, as displaced workers struggle to find new employment. Additionally, there is the risk of social unrest, as large segments of the population may feel left behind by rapid technological advancements.
Another problem is the challenge of retraining workers for new roles. Many displaced workers may lack the skills required for the emerging job markets, particularly in fields like software development, data analysis, and AI maintenance. Governments and educational institutions will need to invest in retraining programs to ensure workers can adapt to these changes.
In conclusion, while computers and automation bring significant benefits in terms of efficiency and cost savings, they also pose challenges in terms of job displacement and social implications. The key to addressing these issues will lie in preparing the workforce for a rapidly changing job landscape through education, retraining, and social support systems.