I received beta access to OpenAI’s GPT-3 playground the other day, so I attempted to have it write an article about “extreme heat vulnerability and smart cities.” I put that prompt into the Essay Outline example, then fed the outline topics into the chat feature, and finally assembled the text below. I selected the outcomes from the chat that best fit the prompt. Sometimes GPT-3 would go on a tangent, at which point I reset the chat window and started over again with the next question.
Parts of the response are objectively wrong, such as the definition of heat vulnerability. Other parts are clumsy, like the concluding header “How the Future of Smart Cities Can Change Our Future". It was also hard to coax lengthy text from the GPT-3 chat playground. Even with the response-length setting tuned up to more than 1,000 characters (about 8 sentences), it only provided a few sentences at most. Finally, it provided contradictory responses to the same prompt when refreshed, as seen in the two responses I included under the “How do smart cities help mitigate extreme heat vulnerability?” heading.
I was curious if GTP-3 could replace a human researcher and author. At this stage, I do not think it is quite there. However, it may have some utility in filling in certain gaps, or being a chat partner to bounce ideas off of.
Without further ado, I present OpenAI’s essay:
Extreme heat in cities is the energy consumption by cooling devices when the temperature is above normal. The cooling systems that use water, chilled water, and chilled air consume more energy than those that use a dry cooler.
There were a number of extreme weather events that impacted the living conditions in major cities. These said events caused a severe decline in air quality, higher mortality rates, higher hospital visits for respiratory and cardiovascular illnesses, higher demand for emergency services, economic loss, and so on.
There are a number of factors that make a city more vulnerable to extreme heat. They are: urban heat-island effect, climate change, less vegetation in the city, air pollution, and lack of city readiness.
In smart cities, cities are designed with intelligent systems and solutions that help reduce energy use and carbon footprint. AI helps manage different parts of a city, such as monitoring air quality, waste water, and traffic patterns. Also, AI can be combined with existing technologies to provide better services in smart cities.
Smart cities have many benefits for the inhabitants of the city. By collecting and analyzing data, smart cities can reduce traffic congestion, guide visitors to tourist locations in the city, and provide a safer environment. To further their goals, smart cities can make use of autonomous vehicles and public transportation to reduce the number of cars on the road. Public programs can be managed by smart cities to support people's education and financial needs. The improved resources made available to citizens can help them live more fulfilling lives.
There is a significant decrease in heat-related deaths and illnesses in smart cities, as they have access to real-time information and alerts about extreme weather events. Cities can thus improve their safety and response to high temperatures, helping to keep infrastructure running. (response 1)
Weather-related deaths and illnesses have decreased significantly in smart cities with Internet of Things technology. However, there is little evidence that nonphysical aspects of smart cities have improved. Researchers are beginning to focus on broader policy implications regarding the quantifiable, nonphysical outcomes of smart city development. (reponse 2)
Smart cities will be a global trend in the future. The development of smart cities is very fast, more and more smart devices are being used to create better living conditions.