ChatGPT is already causing some concerns. The text generator from OpenAI is exceptional at creating text for various prompts quickly, but it may not always produce factually accurate or engaging content. Its versatility is remarkable, but it also has the potential for misuse, even with built-in filters.
It's unsurprising that people have found questionable or even frightening uses for the tool. Anything capable of generating content has the potential to create something harmful, and ChatGPT operates around the clock. Keep in mind that these uses have surfaced during the tool's early stages, before it becomes fully mainstream.
One of the more concerning applications is the creation of malware by ChatGPT. While malware itself is nothing new, ChatGPT's ability to generate it endlessly raises serious security concerns. Infosecurity Magazine reported that "cybersecurity researchers were able to create a polymorphic program that is highly elusive and difficult to detect." In essence, the researchers used ChatGPT to generate malware code and produce variations that are hard to identify or stop.
"In other words, we can mutate the output on a whim, making it unique every time," the researchers told Infosecurity Magazine.
While less alarming, students are using ChatGPT to cheat in school. With a tool capable of generating text on any topic, students are tempted to employ it for academic dishonesty. Professors have reported catching students using ChatGPT to cheat, leading some school districts to ban the app. This trend is likely to persist, potentially leading to AI-assisted learning for students.
While not exactly spam, some individuals are using ChatGPT to have conversations with their Tinder matches, essentially outsourcing their conversational efforts to AI. It's unsettling to think that you might be chatting with an app rather than a genuine potential partner, reducing dating to a growth hacking exercise.
It's more challenging to provide concrete evidence, but it's reasonable to assume that ChatGPT or similar AI writing tools are ideal for phishing attempts. Phishing messages are often easily identifiable due to their poor language, but ChatGPT can rectify these linguistic issues. Experts have suggested that this is a potential use case for the tool.
In an experiment, Mashable asked ChatGPT to improve the language in a real scam email, and not only did it enhance the text in seconds, but in one of the responses, it even started blackmailing the fictional recipient without any request to do so.
Job applications can be an exhausting and sometimes disheartening process. If you're actively seeking employment, you might not be aware that your dream job could be snatched up by an AI app. According to a consultancy firm, applications written by ChatGPT outperformed those written by humans in 80 percent of cases. It's easy to see how ChatGPT could excel at including the keywords that grab recruiters' attention or, more importantly, pass through HR software filters.