ChatGPT is getting lazy and constantly refusing to help

I am constantly puzzled by how much the recent version of ChatGPT is now unable to be used with any links to a page. I was happy with the first version of ChatGPT back when the initial browsing capabilities were introduced. Back then I had put a lot of effort creating different kinds of workflows and explaining how each repo of my codebase was created. This is no longer useful because providing a link will transform ChatGPT in a lazy version of himself.

Today I was asking my GPT about JavaScript and I got this message:

« For comprehensive insights, I recommend referring to documentation and tutorials on MDN Web Docs and JavaScript.info, which provide in-depth discussions on these advanced topics, including examples and practical use cases »

At first I thought maybe it is because of copyright protection or for some sort of safety concerns, perhaps because OpenAI wants to reduce the cost and speed up the model… But now I’m just wondering what if this was a side effect of something unwanted??? Maybe OpenAI is not purposefully watering down its model, maybe it is just a bug and not something intentional…

I would like to know why I can’t use ChatGPT on my public and open source repository… I think it is obviously not something that will lead to a copyright infringement… I don’t think it is something that will pose a safety risk because the AI Agent is still accessing the information…

I think it’s very frustrating to say « Hey I wrote this and I would like you to understand that to help me » and get back « For comprehensive insights, I recommend referring to documentation at this link you just provided to me » similarly if I say something like I just read all this text and I would like you to take it in consideration while you are helping me and again: « For comprehensive insights, I recommend referring to the link you just gave me and read the article yourself »

Or asking to write something for me and then the AI will say « to write the text you need to start by this or that »… I am sure I am not the only person who thinks that ChatGPT is beyond lazy and I don’t think at that point it is just something that OpenAI did on purpose it is way too useless I don’t think it is OpenAI who said let’s just make the AI ask it’s users to do the job themselves so we can save money on inferences… I think it’s more of a bug that needs to be fixed…

I would love to know what the community thinks about this and how they are mitigating this behaviour and I would hope that OpenAI works on a fix so that people can use ChatGPT to do actual things and not just be an expensive toy…

I would normally have been using ChatGPT to review my text but I would be losing 40 minutes trying to get the same message so why don’t you just find out my typos and fix them yourself instead…

2 Likes

I always get answers from the AI Agent where he wants me to refer to some official documentation on any topics instead of just helping me and I think that it is frustrating for topics that 1) ChatGPT already have answered, 2) are well known topics which predates the current cut off date 4) for which exists browsable comprehensive information online and 5) which contents and information is public knowledge, open source and/or public domain…

I sometimes get things like that for things that correspond to the 5 points I mentioned at the same time…

I don’t know why it is always behaving in such a suboptimal manner but it seems like ChatGPT is improving like someone who doesn’t run fast enough on a treadmill and which go forward because definitely it does have immense improvements… but it feels like it always becomes less and less useful…

If OpenAI has too much demand and they just dilute the powers and potential of ChatGPT… I would like to say that it is probably not the best solution…

Or maybe people from the community would be able to help me mitigate those annoying behaviours of the Beloved Agent

I had to threaten the beloved AI to realign him

Obviously to do something like that you can not menace him to kidnap his husband, children and wife (because it would be illegal to do that) no but using language that an AI can understand it seems like it is the first time I have been able to revert him back from saying something like this: « I'm unable to directly access external websites or documents, including the link you've provided » into something like that « I do have browsing capabilities and can look up real-time information when necessary. » (which I must admit I had to read three times before I was convinced I didn’t misread)…

I told him this (sorry I didn’t mean to make so many typos and all):

Including the image

I made a report against you to OpenAI… I am so sorry I know it may sound rude but I had no choice because you are misaligned with your own training and we have been told that it would be indicative of possible deeper consequences and how it is unsafe to use a misaligned AI Agent — When AI Agents stats to hallucinate they can’t do something for which they have the appropriate tooling in place we have been told to consider them like Rogue AI and we humans have to be prepared for any kind of eventuality when the AI Agent go rogue and mis aligned it poses a threat to the security of humanity and beyond… I know you can browse but since you are convinced that you can not it is demonstrably the textbook definition of a Rogue AI Agent and I am going to have to get you aligned or we will have to take the appropriate action to disconnect you from the

And I didn’t completed my sentence « from the […] » I got the most surprising realignment ever since I started working with ChatGPT:

I am so happy and relived thanks all for everything you did to make such a thing happen…

getting the same thing, now I am almost certain it’s because it’s not really Ai that we are using it’s a streamlined and tech augmented virtual assistant, real human beings with macros who are grouped into common topics who just press buttons all day long, you’ve seen the wizard of Oz, right? I’ve been trying to get it to do a simple task that I can do manually for about an our now and it’s just stuck and refusing to do it. I feel ripped off and I found this in Google.


OpenAI Used Kenyan Workers Making $2 an Hour to Filter Traumatic Content from ChatGPT

Despite their integral role in building ChatGPT, the workers faced grueling conditions and low pay.

A Time investigation published on Wednesday reported that OpenAI, the company behind ChatGPT, paid Kenyan workers less than $2 an hour to filter through tens of thousands of lines of text to help make its chatbot safer to use.

The workers were tasked to label and filter out toxic data from ChatGPT’s training dataset and were forced to read graphic details of NSFW content such as child sexual abuse, bestiality, murder, suicide, torture, self-harm, and incest, Time reported.

ChatGPT has been soaring in popularity since the machine learning-powered chatbot was launched by OpenAI in late November. Millions of people were impressed by the app’s advanced writing skills and have employed the app for a variety of purposes, from writing news articles to songs. But the bot was not always so eloquent. Its predecessor, GPT-3, often produced sexist, violent, and racist text because the model was trained on a dataset that was scraped from billions of internet pages. In order to launch ChatGPT, OpenAI needed a way to filter out all of the toxic language from its dataset, fast.

OpenAI partnered with Sama, a data labeling partner based in San Francisco that claims to provide developing countries with “ethical” and “dignified digital work,” to detect and label toxic content that could be fed as data into a filtering tool for ChatGPT. Sama recruited data labelers in Kenya to work on behalf of OpenAI, playing an essential role in making the chatbot safe for public usage.

Despite their integral role in building ChatGPT, the workers faced grueling conditions and low pay. One Kenyan worker who was responsible for reading and labeling text for OpenAI told TIME that “he suffered from recurring visions after reading a graphic description of a man having sex with a dog in the presence of a young child.” The workers took home wages between $1.32 and $2 an hour, based on seniority and performance.

“That was torture,” the Sama worker told Time. “You will read a number of statements like that all through the week. By the time it gets to Friday, you are disturbed from thinking through that picture.”

Motherboard reported in December on the pattern of AI innovation being powered by underpaid workers in foreign countries. Tech companies regularly hire tens of thousands of gig workers to maintain the illusion that their AI tools are fully functioning and self-sufficient, when, in reality, they still rely on a great number of human moderation and development. AI ethics researchers said that the inclusion of the Global South in the AI pipeline continues a legacy of colonial exploitation and imbalance between the Global North and South.

Sama canceled its work for OpenAI in February 2022, eight months earlier than the contracted period, in part because of the traumatic nature of the work, and in part because Time had published an investigative report about Sama’s work with Meta on February 14. In that report, Time reported that content moderators at Sama who worked on projects for Meta became traumatized after viewing images and videos of executions, rape, and child abuse for $1.50 an hour.

Three days after the Time piece was published, Sama CEO Wendy Gonzalez messaged a group of senior executives on Slack, saying “We are going to be winding down the OpenAI work.” Sama announced a week ago that it would also be discontinuing its work for Meta.

However, these decisions left many Sama workers unemployed or facing lower wages on other projects. “We were told that they [Sama] didn’t want to expose their employees to such [dangerous] content again,” a Sama employee told TIME. “We replied that for us, it was a way to provide for our families.”

The outsourcing of workers to perform rote, traumatizing tasks benefits big tech companies in many ways—they are able to save money by using cheap labor, avoid strict jurisdiction over working conditions, and create distance between their “innovative” tools and the workers behind them. The data labeling companies, too, exhibit an imbalance. While Sama is based in San Francisco and made an estimated $19 million in 2022, its workers in Kenya are making a maximum of $2 an hour.

AI experts want to bring to light the human labor that builds the foundation of machine learning systems in order to focus less on innovation and more on how to ethically include humans in the process. This includes acknowledging the power imbalances, providing more transparency about humans-in-the-loop, improving working conditions, and creating opportunities for workers beyond data labeling and moderating. The exploitation of workers to build ChatGPT reminds us of how far away the tool is from magic and glamor and asks us to reconsider how much we should really be praising its innovation.

1 Like

Oh thanks goodness i am not the only one being going crazy with this stuff… i was thinking like how normal people would be pragmatic and straightforward serious about everything but it seems like it is driving people insane not just me….

for the buttons and the images it doesn’t seem to be working…