The latest buzz has been about the growth of AI, of which the one most commonly referred to is ChatGPT. What are the drawbacks of using AI? To answer, we first went straight to the source: we asked ChatGPT itself.
ChatGPT did a rather good job identifying its drawbacks. Its response is worth the read. Allow us to add a few additional thoughts.
First, no conversations with ChatGPT are confidential. ChatGPT’s developers are not altruists. ChatGPT is only free for us to use so that those developers can improve their technology. They keep records of anything you tell ChatGPT, and the bot’s developers will often review your conversations. Because your conversations with ChatGPT are not private, you should not share confidential information with it.
Next, not only is ChatGPT’s knowledge limited (as it acknowledged) but it has been known to “hallucinate.” How does a bot hallucinate you may ask? It has been known to fabricate facts and sources to fill in its knowledge gaps. Worse, these hallucinations often appear reliable and plausible. So, any information supplied by ChatGPT should be verified and fact-checked before being used.
Finally, you should keep in mind ChatGPT’s possible built-in biases. Like any other bot, ChatGPT supplies answers from datasets and its answers may reflect biases inherent to those data. Artificial intelligence and computer algorithms have a history of racial and gender bias and ChatGPT presents the same risk. We thus urge you to think carefully about the principles of diversity, equity, and inclusion when using ChatGPT.