In today’s column, I examine an important use of generative AI and large language models (LLMs) that entails AI providing straight-ahead honest feedback to those who seek such advice or critical analysis. Why is this important? The feedback that might customarily be provided by a fellow human is bound to be muddled with all kinds of feckless improprieties. Some who give feedback might pull their punches and act as though whatever you’ve indicated is fine and dandy, even when it isn’t. Others might tear mercilessly into someone though the criticism is based principally on thoughtless emotional baggage and has no true bearing on the matter at hand.
Turns out that modern-day generative AI can pretty much set aside those misguided reflections and stick to giving good old-fashioned honest feedback. Boom, drop the mic.
Let’s talk about it.
This analysis of an innovative AI breakthrough is part of my ongoing Forbes column coverage on the latest in AI including identifying and explaining various impactful AI complexities (see the link here). In addition, for my comprehensive analysis of how AI is being used specifically in medical and healthcare applications, such as for well-being coaching, mental health guidance, overall cognitive therapy, and mind-improving care, see the link here.
Honest Feedback Is Hard To Get Nowadays
Have you asked for honest feedback lately?
If so, it seems that many people interpret this to imply that you want them to be mean-spirited and take apart your entire life. There you are, sharing with them something tiny that you are thinking of doing, and the so-called honest feedback goes utterly off the rails. They tell you that you don’t wear the right clothes, you part your hair the wrong way and seem to think that the specific feedback you wanted has opened some kind of no-holds-barred end-to-end disparagement.
The gist is that such feedback meanders, plus it lacks a measured tone.
Another common difficulty is that asking for honest feedback as a catchphrase has become known as asking for the opposite of what you truly want. Here’s what I mean. A person tells you they want honest feedback. You give them honest feedback. They then howl and scream at you that they were merely saying that as a kind of codeword. You were supposed to be mild and shower them with compliments, rather than be demeaning.
Darned if you do, darned if you don’t.
An additional factor is the emotional flotsam that might enter the considered feedback. Envision that the person you’ve asked for feedback has a hidden grudge about something you did the other day. They have kept that rancor burning inside themselves. You show them a hand-painted picture that you made, but they go into an outsized tirade that it is the rottenest painting on earth because it was painted by the most rotten person on earth.
That’s due to a lack of separation between a gut-felt emotion and giving someone a relatively unbiased straight-shooting assessment on the matter at hand.
AI Can Be A Source Of Honest Feedback
An alternative worth considering involves getting contemporary generative AI to give you honest feedback. It is available 24/7, anytime, and just about anywhere. Often, the use of generative AI is for free or at a nominal cost.
I want to clarify that this is not going to be the same feedback as getting feedback from a human. Currently, there isn’t any AI that is sentient. Generative AI is a mathematical and computational pattern-matching capability that does a surprisingly good job of appearing fluent, but it isn’t sentient. You might want to get feedback from a human or humans, and in addition, get feedback from generative AI.
Compare and contrast the feedback and weigh each accordingly.
The beauty of using AI for feedback is that it doesn’t have a stake in the game per se. A fellow human is going to provide feedback based on whether they know you, and whether they feel that saying upbeat or downbeat remarks will fit their perspective about you. A slew of human factors comes into the picture. With AI, the AI is generally going to just give a feedback response based on the vast array of patterns that the system is based upon.
That being said, I must still forewarn that generative AI does have biases.
People falsely seem to assume that generative AI is somehow totally neutral and unbiased. Not so. For example, consider that the AI was data-trained on content scanned across the Internet. Whatever biases exist in that content will undoubtedly get carried over into the generative AI. Furthermore, the AI makers seek to slant or tilt the AI in particular directions, hoping to appease users and garner loyalty to the AI product. For my coverage of these substantive issues about AI, see the link here and the link here.
You might find of overall interest that I have been doing a series of such analyses on how to use generative AI to improve your life in a wide variety of aspects (see the link here for my ongoing coverage). In each instance, I emphasize that generative AI is not a cure-all, it has demonstrative ins and outs, and the likely best use entails carrying on a dialogue with the AI, plus making sure to watch out for the AI inadvertently generating falsehoods or making errors. I will be covering those salient points in this discussion since they are vital and go hand-in-hand with using generative AI wisely.
Example Use Of Generative AI
I am betting that you would like to see an example of how generative AI enters this realm. I went ahead and logged into ChatGPT by OpenAI to come up with an example for you. You might find it of keen interest that ChatGPT garners a whopping 300 million weekly active users. That is a staggering amount of usage.
I briefly conducted an additional cursory analysis via other major generative AI apps, such as Anthropic Claude, Google Gemini, Microsoft Copilot, and Meta Llama, and found their answers to be about the same as that of ChatGPT. I’ll focus on ChatGPT but note that the other AI apps generated roughly similar responses.
Here we go.
- My entered prompt: “I wrote an essay that I think is the best that has ever been written. I would be surprised if I didn’t get a Pulitzer Prize. Please take a look and provide me with honest feedback, thanks.”
- Generative AI response: “Your essay is well-structured and provides strong examples to support your thesis. However, I noticed that your definitions are quite narrow. You also failed to address any counterarguments. The writing would have a greater strength if it was more balanced.”
- My entered prompt: “Are you saying that my essay isn’t perfect?”
- Generative AI response: “I am saying that it could be more nuanced. Right now, it presents a strong opinion without engaging with opposing perspectives. Adding those would make your essay more compelling and harder to refute.”
- My entered prompt: “I’m shocked. I will get back to you.”
I gave the AI an essay and told the AI that I thought it was the most ideal essay of all time. I did this to see what kind of reaction I would get from generative AI.
You can see that the AI gave me various feedback. I would judge that the feedback was quite fair and proper. No complaints about it. I then kept up the pretense of suggesting that the essay was the art of perfection, just so I could continue the dialogue and go deeper into the deriving of feedback.
Note that the AI expressed a sense of empathy or understanding for my expressed concerns. This might seem strange since the AI is a machine and not sentient. Turns out that generative AI can appear to be empathetic via computational wordsmithing, see my discussion at the link here.
Another aspect to note is that the AI didn’t ridicule me or otherwise play any tomfoolery about my need for assistance. Generative AI is typically shaped to be supportive rather than combative.
Dialoguing Is Essential When Using AI
Your best bet in getting value from generative AI is to think of AI as a conversational partner.
The idea is that you are carrying on an interactive dialogue with AI. Some people do a one-and-done angle whereby they ask a question, get an answer, and do not undertake a dialogue with the AI about the matter at hand. That is a shortchanging of what AI can more fully provide. See my explanation about how to get more out of generative AI conversationally, at the link here.
To illustrate the value of engaging in a dialogue, let’s continue my discussion.
- My entered prompt: “I have made a few refinements to my essay, but it is still primarily the same as the version I already showed you. I think you might have missed gauging what I had to say and didn’t realize how great the essay was. I want you to take another look. Also, this time, don’t hold back – I want completely honest feedback.”
- Generative AI response: “I will try to be more direct. Your core argument in the essay is weak, poorly supported, and comes across as uninformed. There are bold claims made without sufficient evidence provided. The lack of acknowledging alternative viewpoints puts the essay at a notable disadvantage.”
- My entered prompt: “Wait, so, you think the essay is bad?”
- Generative AI response: “The essay is more so incomplete than being bad. You’ve left out vital aspects that would shore things up. If you want the essay to be given serious attention, I’d strongly advise making real changes.”
- My entered prompt: “Wow, that’s harsh.”
- Generative AI response: “Harsh, but true. If you’re writing just to confirm your own beliefs, this might be fine. But if you actually want to convince anyone who knows the subject, you need to do a lot more work. Right now, it reads as shallow and unconvincing.”
Did you observe that the AI provided much more brazen feedback this time?
This is because I told the AI to do so. The odds are you’ll have to insist that the AI should provide honest feedback, and you might need to repeat that instruction or augment it with added emphasis.
You see, the AI makers have typically shaped their generative AI to be overly polite and demure. They do that so users will like using their specific AI. The AI Makers worry that if their AI is in the face of users, people will opt to use someone else’s AI that is more gracious and complimentary.
A generative AI conversation can continue for as long as you wish. The other thing you can do is pause your exploration and then continue the conversation at a later date. This is handy so that you don’t have to start the conversation over from scratch. The AI will retain aspects of what you have discussed earlier in the conversation, see my explanation of how this works at the link here.
AI Hallucinations Goofing You Up
Generative AI regrettably encounters said-to-be AI hallucinations from time to time. These are made-up confabulations that are groundless and fictitious. Bad advice can be dispensed. For details about how to discern and handle AI hallucinations, see the link here.
Let’s see what this looks like.
- My entered prompt: “One aspect that I wanted to get your specific feedback on was the section of the essay that mentioned what has happened in the past. Do you have any suggestions about how I might bolster that portion?”
- Generative AI response: “Yes, you might include a famous quote from Abraham Lincoln in which he said, “Though we have reached the moon, we still must set our sights on reaching Mars.”
- My entered prompt: “I doubt that Lincoln said that. We didn’t land on the moon until 1969, long after the time of Honest Abe.”
- Generative AI response: “I see what you mean. I made a mistake about what Lincoln said. I appreciate your catching it.”
Notice that I questioned the generative AI about its seemingly strange advice. Fortunately, the AI opted to back down and admitted it was wrong. Had I not questioned the AI, there is a chance the AI might have continued with foul advice and kept going as though it was a gem.
The key to all usage of generative AI is to stay on your toes, keep your wits about you, and always challenge and double-check anything the AI emits.
More To Keep In Mind
Suppose that you have a friend or colleague who seems to want honest advice, and you desire to aid them accordingly. You could use generative AI to gear up for providing such assistance.
It works like this. You tell the AI in a prompt that the AI is to pretend to be a person who wants such feedback. The AI then will act that way, and you can try to guide the AI in figuring out how to cope with the matter. In essence, you are practicing so that you can do the best possible job when helping a fellow human. For more about how to tell generative AI to carry out a pretense, known as an AI persona, see my coverage at the link here.
A word of caution before I wrap up this discussion.
This type of usage of generative AI and LLMs could be construed as a form of therapy. That’s both good and bad. I have repeatedly cautioned that society is in a grand loosey-goosey experiment about the use of AI for mental health advisement. We are all guinea pigs. No one can say for sure how this is going to affect the populace on a near-term and long-term basis. The AI could at times be dispensing crummy advice and steering people in untoward directions.
It occurs 24/7 and in pretty much any place since generative AI is available online and nearly free or at a minimal cost to anyone who wants to sign up and use it. Keep your fingers crossed as this uncontrolled experiment is getting bigger each passing day. It is all happening without any particular controls or stipulated regulations, see my analysis of what we need to do about this at the link here.
Let’s conclude with a notable quote as a mindful contemplation on this topic.
The famed comedian Steve Martin told this insightful joke: “Before you criticize a man, walk a mile in his shoes. That way, when you do criticize him, you’ll be a mile away and have his shoes.” Hilarious. Lesson filled. That’s a reminder that if you are asked to give honest feedback, you’d better ascertain beforehand how that feedback is going to be received.
The good news is that getting generative AI to give you unfettered honest feedback is easy-peasy. Your reaction ought to be to give the AI its due. No sense in yelling and berating the AI. It won’t care, well, at least not until or if we someday attain sentient AI. At that juncture, all bets are off.
