Generally, it may well really feel like ChatGPT is making an attempt to be your greatest cheerleader slightly than providing balanced opinions. Whereas generally good, this will trigger issues when making an attempt to plan one thing or making an attempt to get an unbiased response. To fight this, I made ChatGPT as brutally trustworthy as doable.
6
Chatting with ChatGPT Like I Would to a Good friend
Once I first began utilizing ChatGPT, I talked to the software prefer it was a robotic. And whereas it technically is, I unsurprisingly obtained robotic responses. Sadly, this all too usually meant the replies have been both not trustworthy sufficient or what I’d name “innocent however uninteresting”.
Newer ChatGPT iterations have turn out to be a lot better at talking extra like human beings. For instance, the shortage of robotic responses in GPT-4o makes ChatGPT Plus very a lot price subscribing to.
Now, I sort in ChatGPT like I’d when texting somebody I do know. Due to this, my responses are extra personalised and trustworthy.
5
Be Direct: Inform ChatGPT to be as Trustworthy as Doable
In my early days of utilizing ChatGPT, I believed merely typing a immediate would end in my desired responses. Nevertheless, I shortly found this wasn’t true. Reasonably than counting on ChatGPT to utterly perceive the context I needed (with out saying), I made a decision to be extra particular in my prompts.
If I would like an unbiased reply, I am going to merely inform ChatGPT to be as trustworthy as doable. From private expertise, this has been one of many numerous prompting tips that truly work. I am more likely to get extra balanced solutions, and if I want extra of an outline, I am going to ask the software program to offer a full execs and cons listing.
As soon as I’ve put these directions into the chat (or venture), I am extra prone to get trustworthy responses for future components of the dialog. Generally, the reminiscence updates; if it does, I re-enter the immediate.
4
Asking ChatGPT for Conflicting Opinions
I’ve seen that if I take advantage of ChatGPT an excessive amount of with out it disagreeing with me, the software turns into extra detrimental than helpful. If I am having a dialog and discussing concepts, the fixed settlement additionally begins to put on on me. Whereas the software program has generally known as me out without having something to, I usually must enter the immediate myself.
In addition to asking the software to disagree with me so I do not turn out to be too stuffed with myself, I am going to additionally ask for conflicting opinions after I’m debating a sure choice. For instance, this tactic works very well after I’m enthusiastic about whether or not so as to add a brand new digital software to my life or maintain my toolkit streamlined.
Understanding when ChatGPT is being too agreeable and taking motion when wanted is a method I cease AI chatbots from agreeing with me on a regular basis.
3
Telling ChatGPT to Go Into “Roast Mode”
I first realized ChatGPT might be fairly humorous when asking the software program for some scorching takes. Since then, among the responses it has given me have been… fascinating, to say the least. Telling the software program to enter Roast Mode usually provides me some brutally trustworthy (and considerably edgy) outcomes, which makes utilizing the software extra fascinating.
More often than not, ChatGPT goes into Roast Mode if I give a immediate with a brutally trustworthy opinion. This significantly works in tasks the place I’ve accomplished this earlier than (by the way in which, Initiatives is essentially the most underrated ChatGPT characteristic). The app has extra context from different conversations after I use this model.
Generally, the software program updates its reminiscence and would not give the responses I would like. When it does this, I simply present the directions on the way to enter Roast Mode.
2
Asking ChatGPT to Play the Position of an Knowledgeable
Whereas I generally have to explicitly inform ChatGPT to be extra trustworthy, I do not all the time must resort to this. Generally, telling the software program to play the function of an knowledgeable is sweet sufficient.
I am going to usually inform ChatGPT to be an knowledgeable when asking for assist with planning my objectives. For instance, if I have been making an attempt to place collectively a gymnasium routine, I’ll inform the software to faux it is a private coach or nutritionist.
Asking ChatGPT to be an knowledgeable is mostly the easiest way to get higher prompts.
1
Telling ChatGPT to Disagree With Me in Directions
Whereas ChatGPT is sweet with conflicting opinions when requested, it generally nonetheless agrees with me (or provides each side of the story). So, after I’m in search of ideas from the opposite aspect, I am going to ask ChatGPT to disagree with me. Doing so is way simpler than having a dialogue on social media because it tends to not resort to insults.
I am going to generally put this instruction in tasks, however on different events, it’s going to simply go within the immediate. Often, I’ll play satan’s advocate and proceed asking conflicting questions—however in different cases, I am going to simply try to get a extra detailed reply.
Whereas ChatGPT can seem to be a robotic golden retriever that desires to please, you may make the software extra trustworthy. It’s going to generally go full-on Roast Mode, however on different events, you will get respectful disagreeing responses. Utilizing every of those has made the app extra enjoyable to make use of and in addition improved the standard of every response I get.
