Hello everyone! For my blog post this week, I wanted to talk about the recent big news surrounding AI: the release of Chat GPT-4. I am sure that most of us have already heard the news, but I think it is important to highlight its capabilities that are built upon the previous version, Chat GPT-3.5. The previous version of Chat GPT surprised many users with what it was able to do from solving coding questions to writing blogs. While many people fear that Chat GPT may result in a loss of jobs, some companies like Slack have already begun to incorporate this tool for certain roles to help their business, for example, creating an AI-powered assistant to answer questions that users may have on the company’s website. This post will provide an overview of Chat GPT, what is new with the release of Chat GPT-4, and explain how this tool can change the world.
What is Chat GPT?
Chat GPT is an AI-powered chatbot tool created by OpenAI, a company founded by Sam Altman and Elon Musk. In November 2022, Open AI announced the anticipated release of Chat GPT, which used the GPT-3.5 large language model. Large language models are machine learning models that are trained on significant amounts of data that interact with inputs similar to a human. After its release, this AI tool surprised the public as it allows users to enter prompts directly on the website and generates human-like images or text. Some use cases for Chat GPT include coding with prompts and writing blogs. Chat GPT-3.5 amazes many people with its abilities, but there are some major downfalls to this version of the tool. On Twitter, I have seen many people discussing Chat GPT’s inability to truly write like a human. Most of the reviews reveal that the chatbot has the potential to use information that is incorrect or unreliable, cannot write with the flow and emotional intelligence of a human, and is unable to solve complex math problems. In January, Chat GPT-3.5 took the bar exam and scored in the bottom 10% of test takers. OpenAI looked to address many of these issues with the new release of Chat GPT-4.
What is New with Chat GPT-4?
Last week, OpenAI announced the release of Chat GPT-4. Chat GPT-4 is accessible to those who pay $20/month to become a Chat GPT Plus subscriber. This new version still has some flaws experienced with Chat GPT-3.5, but GPT-4 is trained on more data and has improved slightly on the creativity and bias concerns. Last week, Chat GPT-4 took the bar exam, and the scores were significantly better than the Chat GPT-3.5 exam results in January. Chat GPT-4 was able to score in the top 10% of test takers, a huge improvement from the Chat GPT-3.5 score in the bottom 10% of test takers. Another significant update surrounding Chat GPT-4 is the tool’s ability to receive inputs that include images and answer questions surrounding the image. This ability can be useful for analyzing an image for elements that stand out or explaining the meaning behind certain images like paintings. These new capabilities improve the functionality of Chat GPT, but important flaws from the previous version still exist. Chat GPT-4 still tends to “hallucinate”, or provide incorrect information as if it were a fact. Also, the current version of Chat GPT has been criticized for having bias on Twitter. One example of the chatbot’s bias can be found in the picture of the tweet below when a user asked the AI tool to generate a function based on race and gender to check if someone is a good scientist.
Potential Effects of Chat GPT:
Although Chat GPT has many flaws, this AI tool is causing many people to fear its potential to change the world. Chat GPT has already impacted the education system due to students using the tool to complete their assignments. Teachers are adding class policies that prohibit Chat GPT in their classes, and the AI tool is already banned in New York City schools. With the chatbot already forcing people to adjust, it is important to consider how it will impact the world as it continues to improve. One of the biggest concerns is Chat GPT’s potential effect on the workforce. Currently, the AI tool can perform many tasks that people are hired for including journalism and software engineering. As Chat GPT continues to develop and is trained on more data, there is a serious concern surrounding the loss of jobs. Large corporations like Microsoft and Amazon have shed 10,000 employees and 9,000 employees respectively, and Microsoft announced after the layoff that the company will make a multi-billion dollar investment in OpenAI. Other potential effects of the chatbot include students losing the ability to use critical thinking skills due to their reliance on the AI tool to complete their assignments and people viewing Chat GPT as credible when they are unaware of the biases in the data used to train it.
Final Thoughts:
I think it will be both exciting and scary to see how Chat GPT develops as it receives more investments and users. This tool has many excellent features, but I think there need to be regulations in place to ensure that people will be able to retain their jobs and the tool is not being used for bad purposes. Without any regulations, we could potentially see more people lose their jobs and become exposed to the biases in ChatGPT. Also, the lack of regulations will result in an increase in the use of Chat GPT by scammers for phishing attacks like email scams. In class, we discussed how employees will lose their jobs to the employees who incorporate Chat GPT into their work, but I do not think this will apply in 20+ years as technology evolves.
How big of an impact do you think Chat GPT will have on the world?
Thank you for reading!
Spencer Myers