Privacy with ChatGPT
Should you rethink your questions?
Those questions or curiosities we allow to blow in the wind without an answer because we’re afraid to look or sound stupid or expose intellectual weaknesses can now be asked or explored with the non-judgmental ChatGPT. But should we throw caution to the wind and ask anything? With all certainty an answer will be received. It may be a gentle “I’m sorry, but as an AI model I am not able to…” response but at least your encrypted query was posed to a very intelligent AI entity and within the confines of a semi-private chat session.
ChatGPT does not store your questions for regular users to see or shared with a third-party though your questions are cached within your current chat session. This is so questions and answers can be built upon in the same way we converse with people. When the session is over (by clicking New Chat), the cache is cleared and a new ChatGPT session will have no recollection of past questions. This can be checked by asking for the number of questions asked in the current session.
Using a properly configured browser, a VPN connection and current anti-virus scanning software adds additional layers of security in regards to improving privacy on your end.
Interaction
Good questions are at the heart of a good ChatGPT response but users are advised not to divulge personal information.
I was surprised it was able to answer queries introspectively. For example, I asked, “Would you say our interaction is private?” and that’s how I got the material for the points I’ve made earlier. Below is the ChatGPT response:
Accuracy
Of the hundred or so questions I’ve asked I can’t say any were answered incorrectly or out of context and believe me I was looking for an obvious blunder but so far it has responded with plausible answers to my questions that ranged from the simple to the subject matter expert level. Then again, I made sure to include sufficient language and keywords to help gauge the conversation given the nuances of the English language. With any technology, I find it helps to be aware it’s a tool made by us, to help us, so maybe help it help us?
When I learned something new from ChatGPT, a separate resource was used to expound or corroborate the point but it’s important to note the creators have said the tool may provide inaccurate answers where it seems to make up responses. So, how do you know to trust it? Like anything else, get a second opinion.
Usage
As a futurist, I won’t say I have found a personal assistant or the robotic, low drama friend I’ve wished for but I’m not hesitant to ask the non-personal questions I’ve so often kept to myself or promised to research at some other time in the future. My to-do list of questions were dramatically shortened by spending time with ChatGPT that has unexpectedly gifted our sessions with another growing to-do list but with actionable next steps for me to take. That’s progress!
Perhaps this is cutting edge technology where for now we are closer than ever to being intellectually side by side, neck and neck with AI. Wherever this technology evolves will be new frontier but for now conversations can be had in a way that simulates conversing with a brilliant mind. To an extent, its been trained on all we’ve done and contributed to the human collective. Filtered datasets from the internet, books, articles and so forth help inform ChatGPT along with the genius of the data scientists, computer engineers and generosity of its sponsors. What a leap the way it’s helping write programming code today.
Assessment
There are personal benefits to be gained with this technology that’s similar to the benefits cell phones introduced not too long ago. I have no reservations in ditching my fragile, easily stolen cell phone in exchange for an empowered AI-based GPT (Generative Pre-training Transformer) that’s at my beck and call. But today, just for the asking, I now have a travel itinerary with more details than I would’ve personally included for the road trip I’ve put off for years. A clunky poem was reworded to be more visual and imaginative. Solutions for fixing broken things on the handyman list now have an epoxy solution I can do myself, thanks to ChatGPT.
I hope ChatGPT and other AI models like it evolve in an anthropomorphic direction, more human-like. Who doesn’t admire a patient teacher who knows how to break down complex topics and is not bothered when asked to simplify even further or provide other examples for clarity? Best of all, I’m not intimidated to ask questions in a chat window I wouldn’t dare give voice to (at worse, only the ChatGPT engineers review the conversations for tuning purposes, per their General FAQ published about a week ago). When the question cannot be answered, ChatGPT lets me down gently with the canned “I’m sorry, but as an AI model I am not …” response. Sure, for now.
Synopsis
The fact that chat sessions are semi-private where personal data should not be shared but the non-personal questions we’ve held unto out of embarrassment or the need for an objective conversant, is a good enough reason to free your mind and ask away. For all you know your question could be valid or perhaps just a little ahead of its time. Besides, if you got carried away in a riveting and engrossing conversation with your ChatGPT and would like to delete your data, the site says your conversations are processed anonymously and are not linked to your personal identity. But, don’t take my word for it, ask ChatGPT, then review the platform’s Terms of Use and Privacy Policy.
