Userba011d64_201/istockphoto
As more and more of us are realizing that ChatGPT is a pretty handy (and terrifying) helper, we should know, too, that it can't do everything. There are some tasks AI will wiff completely, so don't even try them — and if you do, be prepared to face the consequences.
While ChatGPT can do wonders with your resume and cover letter, sending a heartfelt, personalized sympathy note may be a bridge too far. While the app can bang out a too long and robotic message to a co-worker, anything more sincere is a reach. AI hasn't mastered human empathy and doesn't know the intricacies about the person grieving that you do, so just do it yourself.
While students everywhere rejoiced when they discovered AI could write essays, teachers took that same moment to find apps that can catch fraudsters. Both Turnitin and Grammarly can detect AI-generated text, and Open AI even created its own tool, OpenAI APL Key, to sniff out AI-generated content. While ChatGPT is a good starting point for research, save yourself the hassle and write the actual paper yourself.
While ChatGPT can generate a screenplay with the characters and plot you choose, don't expect it to be, um, good. While striking screenwriters are demanding guardrails on AI and directors are already wondering if they can get a shootable script out of the app, for the time being, the app is pretty much barfing out hackneyed drivel we don't need.
If you send ChatGPT a query about anything happening since 2021, you'll get this response: "I apologize, but as an AI language model, I don't have real-time information or the ability to predict the future beyond my knowledge cutoff in September 2021 ... It's best to refer to the latest news sources or reliable websites for up-to-date information about the year 2023." So, ChatGPT doesn't know current events.
While people have asked ChatGPT for pep talks and encouragement, looking to artificial intelligence for help with serious issues isn't a great idea — and may even be unsafe. While a therapist must hew to certain guidelines to stay licensed, AI has no such limitations. While you might get some good advice, even ChatGPT says it "may occasionally generate incorrect information,” or “may occasionally produce harmful instructions or biased content.” While AI may offer some surface assistance, don't lean on it for serious matters.