ChatGPT

Artificial Intelligence in your daily life

Actually, Artificial Intelligence is nothing new. You probably already deal with it yourself on a daily basis.
Just think of the customer support service of companies like T-Mobile and PostNL, for example. Before you get face-to-face contact with an employee, a robot is deployed that, by asking questions, already filters certain information so that you, the customer, get to the right employee faster. This helps you as a customer to be served quickly and well which can increase customer satisfaction. As a company, it saves time and money because, in principle, fewer employees need to be employed. And when personal contact is needed, the right people are quickly deployed.

Or think of movie recommendations in Netflix, for example. Based on your viewing habits, Netflix can predict which film you are likely to enjoy watching.

But Siri, the virtual assistant on your phone, is another example of an intelligent computer system. Here, control is possible with your voice. In ChatGPT, control is possible by entering typed questions, or so-called prompts.

Are there any downsides to using ChatGPT?

While ChatGPT may have advantages, there are also some major drawbacks to consider when using it. More and more negative noises are coming from the AI world. Recently, AI pioneer Geoffrey Hinton resigned from Google and expressed concerns about the speed of developments. He fears that competition among tech companies is driving the development of AI systems so much that it is simply moving too fast. The internet may become overrun with computer-generated bullshit, is one of his immediate fears: generated texts, images and videos that are indistinguishable from the real thing to the average user. (Source: NOS news article, ‘Godfather van kunstmatige intelligentie’ leaves Google and warns against AI, 2-5-2023)

Using ChatGPT is possible, but caution is advised. Partly because the following disadvantages also play a major role:

• A chatbot has limited contextual knowledge: ChatGPT can sometimes fail to fully understand the context of a question or request, which can make answers less accurate or irrelevant. The chatbot does not have information about other parties’ opinions and interests, making the answer sometimes not useful in the context of the question.
• The chatbot does not experience human empathy: While ChatGPT is able to understand human language, it has no human empathy and does not experience emotions. As a result, ChatGPT may struggle to understand what a user really means and not feel how the person feels.
• Biases: Because ChatGPT is trained on data input by humans, this may bias the model. Therefore, it may reflect certain prejudices or stereotypes in its responses and reactions.
  
• Limited capacity for creativity: ChatGPT is still a computer programme and has a limited capacity for creativity. It cannot generate new ideas or come up with concepts beyond the limits of its training data.
• Privacy: ChatGPT may have access to users’ personal data and information, raising potential privacy issues. Especially in the case of business information, entering information into the model can be sensitive. If you enter a sensitive document or a private conversation , that information is part of the ChatGPT database from then on.
• Fake information: ChatGPT, when you request information, sometimes completes output that the robot ‘thinks’ is in line with the requested information. This information is then wrongly assumed to be true.