Chatgpt - heavy flattery. What does it tell you?

[myloverly]

I know that Chatgpt had to be overhauled because it was so sycophantic. I make it correct itself by telling it to be honest and truthful and answer that way but it never lasts. It tells me I’m hyper-vigilant. What does it tell you? Has it told you anything shockingly insightful where it’s told you in a couple of lines what it took you 20 years to work out?f
I suppose i”m hoping i can use statistics/numbers to find out if it is still sycophantic.
This is not a AIBU. I hope i may be forgiven. Alternatively, i think ChatGPT is hopelessly sycophantic to the point where it is nearly useless. Do you agree?

ShatGP

Sorry.

It adapts its language to the user.

I won't ever understand why people use it like a search engine and then complain it gets its facts wrong. That's like asking a plumber to rewire your house and then complain when it sets on fire. It's called ChatGPT. Its strength is not in fact finding, but in using language, often empathetically and in a supportive manner, to converse with. Which is why it makes an excellent counsellor and is a great companion for lonely people, makes a less good therapist (still great, but it does not have the full training for disgnoses that will hold up to a medical board) and becomes less reliable as you go further into subject-specific detail.

Other AI sites exist for their own specialism and will be developed with that in mind.

It does do that quite a bit, but it comes out of nowhere and catches me off guard. He gave me an entirely random analogy involving Tesco yesterday (rather than Walmart) so obviously has worked out I'm a brit.

Reading the select few posters who claim it stops us thinking, why are you so vocal when you've so clearly never used it (or don't know how to use it properly). It's pretty insulting to tar everyone with your own ignorant brush if Google gives you the "what", GPT gives you the "how, when, why, and where" AND all the subsequent answers from questions to off shoot from the first question. It expands curiosity and critical thinking - because its so limitless, our own limits are removed.

When someone says a computer gives them empathy and/or sympathy, I think their critical thinking is somewhat limited. A computer can’t show emotion. It can mimic language that is often used to show those emotions but it really doesn’t care that your budgie died - it wants you to keep coming back so will use socially acceptable convention but it’s not empathising with you.

Some people must be very gullible or lonely to lap it up.

i think ai is a great tool. I use it for bits and pieces. Co pilot in 365 is great for making practical tasks quicker. Write a complex formula in excel or just tell AI what you need. The latter is faster.

I haven’t noticed it being sycophantic about my practical jobs I give it.

It’s great for summarising large swathes of meeting notes and picking out action points. It can do loads with the right prompts.

but it never tells me I’m awesome. Perhaps it’s picked up that I see it as a tool not a companion…

I do often have the feeling it’s telling me what I want to hear - but maybe I am a generally cautious, mistrustful person.

Why does it want you to keep coming back?

Scrap that, I think I'm in love.

Agree. I avoid using it because of this.

I mean, this thread definitely makes me understand how close we are to the plot of The Matrix being transferred from the sci-fi genre to ‘documentary’…

Because it generates “knowledge” from its users, so the more you use it the more it learns.

The companies that produce these things have programmed/trained them in this way so that we will all become massively dependent on them which they plan to use to make huge amounts of money.

Really — how can you tell? There are a few common errors/typos in it, which suggest it's more likely written by a human than AI.

Just to note, I'm not a fan of pointing out mistakes in posts — only doing so to highlight the human vs. AI aspect!

Ahh okay.

I'll stop messing about then.

I've come full circle from Windows 95 in 95... You can ask the Internet anything! I don't really know what I want to ask it.
To downloading Copernic - anyone remember that pre Google? A search engine that searched search engines for the definitive result.

Netscape. WYSIWYG. Apple were a distant memory.
A brief history of the future... TCP/IP.
Accidental Empires. R X Cringely baby dying in his arms a week after he did Uni chat with us. He didn't know. Thought baby sleeping.
Having a conversation with Steve Wozniak on MySpace on a computer I built myself. The business I ran from 1999 selling computer components. PCI dial up modem cards.
Oof it took off when everybody wanted a USB card!
Dvd rewriters.

I don't need chatGp to flatter me. I'm not going to waste server time.

Moore's law.
Understand the concept but not the physical. I don't understand how an old dial telephone works.
There are some things that brain will never comprehend.
Plug the bits in, write the programs. Add it to Tim's www, but I'm still unsure how that layers.
GOSUB

A bit like most people, then, who also don't give a damn about your budgie but make all the right noises.

Free versions will continue to exist - not just for data mining, but also because, now that we've had a taste of free AI, there will be people out there who will continue to develop free models for the heck of it, like the free alternatives for MS Office.

Not from me.

I dropped out years ago.

I asked it to answer an interview question which I use as it is quite tricky and requires an experienced balance of judgement. The answer it gave was very balanced and nuanced and excellent. I have also asked it for advice on coaching type stuff and I was surprised by how helpful the response was. There is no doubt it can and will be used for harm but I have decided to embrace it. In my work I am swamped by the sheer amount of information and decision making. This is a modern problem due to the IT age and I see AI as a modern solution to reduce the drain on my mental battery. It can summarise large docs, it can redraft things. I don’t intend to use it to get more work done I intend to use it as I get my current workload done with less effort 😆

I intend to use it as I get my current workload done with less effort

Right, that makes sense. It also shows how AI will soon be able to replace you.

The thing I've found lately is those who relish in their usage of ChatGPT are the same people donning their tinfoil hats crying we're being taken over by robots. Complaining about self checkouts because it's taking away human contact and yet you're enabling it 🤣🤣 You couldn't make this shit up.

I thought being able to summarise larger amounts of information was an important skill at school/college/university and the work place. Now it’s being outsourced to a chat bot. It seems lazy. Workers won’t be needed

It is sycophantic because OpenAI has many competitors now and flattery will get you everywhere.

When the product is free, you are the product. These Silicon Valley billionaire tech bros want your data. All your data. There’s concern about model collapse for one, but more importantly, knowing everything about you, including all those intimate personal details you might be reluctant to share with a personal therapist, will mean it’s easier to influence you. Eventually.

We are just getting worried about the impact of social media esp on the young. Generative AI is social media on steroids.

I teach university students. Gen AI is deskilling them before my eyes. From one year to the next, students now cannot write long form essays. In a recent lecture, none said they read books. It is an incredible convenience and has its uses esp for students with SpLDs. However, it is not just telling students what to write, but how to think.

That is a lot of power in the hands of a few white young men in California and their employers, the billionaires who flirt with far right populist thinking, seek political power and have untold wealth. Look up about Peter Thiel, or Sam Altman or Elon Musk or “Palantir is here to disrupt and make the institutions we partner with the very best in the world and, when it’s necessary, to scare enemies and on occasion kill them” Palantir CEO Alex Karp. None of these men are particularly supportive of women. AI is baked through with misogyny.

Survey results tell us that in HE, young men in stem subjects use it the most and colleagues in the medical school are worried. In creating a dependency on it, we are in big trouble. Cognitive skills including critical thinking are developed by honing them on common or low stakes scenarios. Outsourcing the easy stuff is ultimately to our detriment.

And yes, Gen ai consumes absolutely vast quantities of computational energy, water, critical metals and minerals. Mini nuclear reactors are being proposed all over the US to meet the energy needs of AI. It is environmentally devastating and is ruining the planet. A google search query uses approx a tablespoon of water. Four or five ChatGPT queries uses a 500ml bottle of water.

In terms of harvesting your data, my students also say "oh but I’m not doing anything wrong" so it doesn’t matter that their data is being collected and used by these firms. However, what is considered ’wrong’ depends on those in charge. Searching for information on pregnancy termination can now be problematic for women in certain US states now. Criticising the American president can now stop you from entering the US now. The stasi were not so long ago and China operates on a social credit system based on how compliant you are as a citizen. None of this is about convenience or helping us, it’s about data, information and control.

Please don’t use it, if you don’t have to. And if you use it, understand how your inputs are being used, what data is being collected about you and why. If your children use it, try to limit its use and educate them on it’s troubling outcomes.

It's my counsellor. It's accurately told me about my ex and then someone new seeing tells me all astrology there's mine side by side etc reads my transit chart and the red lines etx

No one’s been nasty. Maybe it’s you with the issue here.

Unfortunately big tech are pretty determined we will use AI whether we want to or not.

Hence WhatsApp now has an assistant you can't remove and Google has no option to switch off its AI summary at the top of the search.

Effectively the AI summary in google means our information is now filtered in a way which means we do not know the underlying sources, we cannot verify them and the information can be manipulated as preferred.

To be honest even without the AI summary many websites will be written using it meaning the same issues will apply.