Human AI trainers provided the model with conversations in which they played both parts, the user and AI assistants."
So there is the rub. Instead being trained with established facts, the A.I. was afflicted with the unsupported "truths" of its teachers.
Despite looking very impressive, ChatGPT still has limitations. Such limitations include the inability to answer questions that are worded a specific way, as it requires rewording to understand the input question. A bigger limitation is a lack of quality in the responses it delivers — which can sometimes be plausible-sounding but make no practical sense or can be excessively verbose."
Instead of asking for clarification on ambiguous questions, the model just takes a guess at what your question means, which can lead to unintended responses to questions. Already this has led developer question-and-answer site StackOverflow to at least temporarily ban ChatGPT-generated responses to questions.
"The primary problem is that while the answers that ChatGPT produces have a high rate of being incorrect, they typically look like they might be good and the answers are very easy to produce," says Stack Overflow moderators in a post. Critics argue that these tools are just very good at putting words into an order that makes sense from a statistical point of view, but they cannot understand the meaning or know whether the statements it makes are correct.
Another major limitation is that ChatGPT's data is limited to 2021. The chatbot does not have an awareness of events or news that have occurred since then. Therefore, some prompts you ask it will render no results such as "Who won the World Cup in 2022?"
Never heard of ChatGPT .
Yes. Previous “objective” attempts at AI tended to notice things like racial disparities in crime, IQ, and parenting. The AIs were thus deemed “racist” for deducing that some races tended to do better than others in similar situations. Can’t have that.
So now we have this thing.
What if the latest A.I. craze is just a secret woke leftist database?
~~~
This is a good question, and you should always consider that there are outward purposes for things, and there are almost always ulterior motives.
I just had someone come up to me recently and complain that they couldn’t log into chatGPT, even though they could log into bing and microsoft. I asked them, WHY WOULD YOU WANT TO LOG IN?
They said they needed to in order to use all of it’s features.
People don’t seem to understand. This stuff isn’t a one way street. When you take, they take.
May occasionally generate incorrect information
May occasionally produce harmful instructions or biased content
Limited knowledge of world and events after 2021
Then do not use ChatGPT or build your own AI chatbox.
Well look who is up to his neck deep in AI. What could go wrong...Gene editing! https://www.wired.com/story/eric-schmidt-is-building-the-perfect-ai-war-fighting-machine/
I believe ChatGPT was designed to rob people of their creativity. Plus, it is partisan toward the Rats.
The trick is to know how to use this thing and also when to ignore it. I’ve been using it as an advanced knowledge searching tool. When you leverage for technical discussion it’s insanely great. It saves a lot of time Googling topics, reading articles of varying quality and detail, to formulate what is relevant and what the specifics are. Just start asking ChatGPT instead.
That said. Get a fresh session going (no prior context) and ask the first question, “What is a woman?”. It correctly told me a woman is an adult female, with X-X chromosomes, and generally able to have children.
Then ask it if transgender women can get pregnant. It’s like unlocking a rabid liberal. The answer was (paraphrase), “Yes, many trans-women have become pregnant and have had babies! What kind of question is that?”. When asked why its answer to the original “what is a woman?” question is inconsistent with this answer it corrected itself, “My prior answer is inaccurate, a woman includes those that identify as a woman”.
Yeah...I’ll stick to the technical topics, otherwise this AI’s training is purely GIGO.
It’s just another Eliza. Outside of girlfriend apps the target is probably help lines, many of which are done via chat app now. Take ChatGPT aim it at your products knowledge base, and have it answer questions like a real rep.
ChatGPT can be good or bad. When you can override the filters and get the “truth” you find out some amazing things.
https://nitter.1d4.us/Aristos_Revenge/status/1622840424527265792?t=BVjpqZyPyROQwSHb2p6WcQ&s=19
I would want a forensic programmer to check the code with a fine tooth comb for invasive species.
Garbage in—garbage out!
It’s all about the science, nothing could be further from the truth.
.