With the arrival of AI tools like ChatGPT, many questions traditionally posted on platforms like ResearchGate are increasingly becoming redundant. This is particularly true for basic queries—such as definitions, literature summaries, referencing styles, statistical help, or academic writing support—which ChatGPT can now address instantly and accurately. Researchers no longer need to wait hours or days for responses when AI can deliver coherent and helpful answers in seconds.

However, this doesn’t mean ResearchGate is obsolete. The platform still holds a unique value in the academic and scientific community. ResearchGate enables direct interaction with field-specific experts, peer feedback on unpublished work, clarification of nuanced research problems, and collaborative opportunities—all of which lie beyond ChatGPT’s current capabilities. It supports discussions rooted in real-world research experiences, something that AI, trained only on existing data, cannot fully replicate.

In essence, ChatGPT is highly efficient for handling routine, general, and technical support questions. In contrast, ResearchGate is evolving into a space more suited for in-depth discussions, expert validation, networking, and cutting-edge research inquiries. Rather than replacing ResearchGate, ChatGPT is redefining its role—shifting it toward more valuable, higher-level academic engagement.

No, asking questions on RG is not totally senseless. You can still take their opinions, discuss them with ChatGPT and learn from the mistakes. RG is very good as an inspiration for questions. Otherwise noone would know what to ask.

This answer was created without ChatGPT. Please check the information.

Paul Tromp Varies. AI is highly likely to give you very generic answer. If that's good enough for that particular question, then it's sufficient. If you have a quite tricky question requiring making non-obvious connections or detecting that on some particular topic dominating view is built on really shaky foundations, then AI would no way detect that by default.

At least once AI gave me answer that indeed must have been what has been the dominating position in its training data, but at the moment of answering it was no longer up to date. It wasn't even training cut-off, but rather old sources completely outnumbering the ones mentioning new findings.

Carmen Wrede "This answer was created without ChatGPT. Please check the information."

Grok, is that true? :D

Marcin Piotr Walkowiak "If you have a quite tricky question requiring making non-obvious connections or detecting that on some particular topic dominating view is built on really shaky foundations, then AI would no way detect that by default."

You mean, just like most people don’t realize that the F-scale has been used against them... so instead of becoming more free and democratic, they’re being subtly brainwashed? I guess everyone has their blind spots, right?

To further clarify Carmen Wrede’s post to readers who have never heard of the F-scale: the “F” stands for Fascism and this scale was intended to detect latent authoritarian attitudes that could support fascist or anti-democratic regimes.

Carmen Wrede "You mean, just like most people don’t realize that the F-scale has been used against them..."

In which way? The whole scale looked a bit odd as it captures at least proxy for 2 out of big 5 factors of personality (openness and conscientious which indeed are correlated with right-left political axis in general), some culture-dependent spirituality stuff with and a few very context dependent, governance questions. It really looked as intended to detect one very particular strain of authoritarianism in particular era while not only failing detecting authoritarianism in general, but even failed to detect communist leanings from that time.

Though the interesting part starts when take it face value - due discourse window shifting quite a lot most of modern populist-right or far-right are going to score on that scale around '50s centrist (or center-left), thus can officially claim not being fascist by test on which quite many are staking their entire reputation to be valid.

Marcin Piotr Walkowiak "In which way?"

What do I know. Could be anything that causes blind spots, like inattentional blindness. I guess if you have a quite complex situation and it requires for people to make connections that are not obvious or people need to detect that a particular topic's dominating view is built on a really shaky foundation... then most people would have no way to detect by default that they are fed with information that have a certain tendency on the F-scale...

Carmen Wrede "then most people would have no way to detect by default that they are fed with information that have a certain tendency on the F-scale..."

While I don't see reasons to take this scale especially seriously (thus I can't see as shortcoming if anyone is not detecting how to score impact of particular message on that particular scale), wasn't mass media and other transmitters of establishment ideology in recent years pushing (or at least paying lip service) for something clearly on low end of that particular scale?

Paul Tromp That was the claimed intention. Setting aside mentioned methodology issues or the issue that one really should take more seriously psychological constructs that are derived in more stringent way through PCA, like IQ/g-factor or Big 5, there is one even more serious problem. It all clings on very convenient assumption that dominant factor are some psychological issues not rational (or bounded rational to be precise) decision. One could instead suspect a bit different dynamics - that people instead of dreaming of some strong guy in charge, would rather have some representative government. However, in case of that government persistently fail, at some point one would be more interested in anyone able to restore order, while sufficiently dysfunctional democracy due to not reflecting will of the people wouldn't be even deemed as something democratic. The second approach would be very inconvenient among so called anti-fasci, as destabilization that they tend to create seem to be key in such shift in public opinion. Moreover, some bounded rationality makes more sense when looking at interwar Europe, where quite many democracies folded and most of subsequent authoritarian regimes maintained rather boring policies.

Marcin Piotr Walkowiak Of course you don't. You are a medical doctor. Asking a medical doctor on that topic would be as if asking ChatGPT about certain topics it cannot provide answer for either. So in both cases when it comes to certain questions it seems senseless either to consult ChatGPT or RG. ChatGPT doesn't know everything. But truth is: Here it is the same... People do not know everything and yet everone acts as if. Now the difference between people here and ChatGPT is, that ChatGPT is fed with approximately 12 Millions PDF's and you just have to ask the right question. Here you can ask the right question and still get wrong answers based on taste, on personal opinions and even on political agenda. Some people really do not act very different than mass media. Do you have any idea that it is like an ongoing joke here for years that people still insist that Einstein was wrong? Or that they really, really want to prove an aether? Of course always based on refined opinions. The problems that come with mass media seem to apply here too. And yet this is not mass media... Still the opinions get shifted. You might think this is just a problem the average person has to deal with, when in fact a lot more people are already under the same spell... because as Wittgenstein said, the grandfather of Kybernetics: Language is bewitchment of the mind. The F-scale was only a first try to see what happens when you computate a lot of opinions.

Carmen Wrede "Of course you don't. You are a medical doctor." That's flattering, but I'm not a MD but a biostatician on public health with a quite multidisciplinary background. (For example I did publish studies on which factors, including political ideology, predict vaccine hesitancy, so I dealt with political ideology in more stringent fashion) Though I'm a bit surprised that you expect others to have very narrow areas of interest as you have in your portfolio so unique combination of literature and physics.

"Here you can ask the right question and still get wrong answers based on taste, on personal opinions and even on political agenda." Chatbot could also provide you with proper ideological agenda - no questions on what Chinese Communist Party would deem as threat for social harmony on DeepSeek, while answers so politically correct that approaching parody would be provided by Claude.

"The problems that come with mass media seem to apply here too. And yet this is not mass media... Still the opinions get shifted." We're observing how mass media lost it sway on political opinions which used to enforce some very narrow discourse window. Right now the discourse window starts getting wide and... wild.

"Chatbot could also provide you with proper ideological agenda"

Nobody here needs a chatbot for ideological agenda. There are enough humans who do the job better than a chatbot. Indeed chatbots learn from them.

@Rinsha K.A I have to disagree with your assertion that ChatGPT can now address referencing instantly and accurately. Hopefully it is now better at it than a year ago, but when I asked chatgpt to flag and correct referencing mistakes in a somewhat sloppy and messy reference list that needed to be transformed into a nice and neat Chicago style reference list, it was unable to spot and correct all the mistakes.

@Marcin Piotr Walkowiak what do you mean with “some culture dependent spirituality stuff”?

Paul Tromp Assuming that I found the right variant:

"[5] Science has its place, but there are many important things that can never be understood by the human mind." (should be most likely intended to be some spirituality but theoretically catches also issues of cognitive limits of our specie)

"[6] Every person should have complete faith in some supernatural power whose decisions he obeys without question."

(Answer depends level of religiosity within society, in highly secular one there would be resound no, even in case of de facto worshiping "dear leader")

@Marcin Piotr Walkowiak In all honesty, I have no idea where you are talking about. Let’s keep this discussion on-topic and related to my original discussion statement.

Seems that Grok's update just made redundant even internet trolls, so humans could concentrate on more meaningful tasks. ;) In spite of some media claims of it being far-right, in Polish cluster of X it appeared neither to spare our populist right nor far right, so rather roasts everyone overboard.

(Turned off some guardrails? Optimized it for max engagement and it delivered?)

Though looking beyond this update blunder, Grok tends so far to be less politically correct, thus useful in case of anyone needing something not following blindly fashionable ideological mindset.

Side note:

Paul Tromp "I have no idea where you are talking about." You asked me about my part of answer on why I'm skeptical of F-scale.

@Marcin Piotr Walkowiak I have no clue who Grok is, and I asked you about what you mean with “some culture dependent spirituality stuff”.

Paul Tromp "I have no clue who Grok is"

That's quite surprising statement for someone who is discussing impact of AI. Grok is one those frontier model AIs, this particular one is trained by Musk's X/Twitter.

Thats true .we write our answers but we are at least checking it with AI..

By unconsciously, We know AI is better than us.😖