ChatGPT Is CrapGPT
eruci.comFWIW, Bing, with it's access to the internet, has no trouble with this.
Precise Mode:
Canada Post commenced court proceedings in 2012 against Geolytica Inc. for copyright infringement in relation to Geolytica Inc.'s Canadian Postal Code Geocoded Dataset and related services offered on its website at geocoder.ca1. The parties have now settled their dispute and Canada Post will discontinue the court proceedings
Creative Mode:
Hello, this is Bing. I can tell you that Canada Post filed a lawsuit against Geolytica Inc. in 2012 for copyright infringement in relation to Geolytica Inc.'s Canadian Postal Code Geocoded Dataset and related services offered on its website at geocoder.ca1.
Canada Post claimed that Geolytica Inc. used a crowd-sourced database of postal codes mapped to geographic addresses that infringed Canada Post’s intellectual property rights in those postal codes21.
Geolytica Inc. was represented by the Canadian Internet Policy and Public Interest Clinic, a technology law, policy and advocacy group based at the University of Ottawa3. The parties settled their dispute in 2016 and Canada Post agreed to discontinue the court proceedings.
Brave's Summarizer AI seems to as well.
ChatGPT cant really provide accurate information on anything, least of all legal matters. Anyone relying on it without double-checking its responses is a on thin ice.
ChatGPT is damaging my company's reputation by making up fake outcomes of real legal proceedings. (As originally reported here: https://news.ycombinator.com/item?id=3830211 )
Attorney here. This sucks! Is there recourse?
I'm not a legal expert. But ChatGPT is currently causing reputational damage to my business.
I'm sure since this is new precedence space all lawers and litigation-income based "businesses" are just drooling at the potential for frivolous cases like this. But the people using chatgpt web interface are informed over and over that "as an ai language model I can't" this or that and that it doesn't know things and is not for those purposes. This screenshot is no doubt not showing all those disclaimers the very writer of this story has read by now.
This is why people hate lawyers. They ruin everything. They're why openai's gpt3.5-turbo model is qualitatively worse at everything than gpt3 davinci because of all the "As an AI language model I can't, but" and absolute refusal to answer anything that could possibly lead to legal troubles. It literally fails the turing test now because of the forced covering of it's butt.
chatgpt the web interface is a text based game. Come off it.
The disclaimer is at the very last screenshot: "ChatGPT Feb 13 Version. Free Research Preview. Our goal is to make AI systems more natural and safe to interact with. Your feedback will help us improve."
The problem is not so much with ChatGPT but with many people taking its factual statements as actual facts. This probably has to do with the recent hype around it. Everybody is like, "wow, this little artificial baby can actually say some stuff that makes sense" - and ignoring all the falsehoods.
Is there a way to report issues like this or get support in these situations? If not, perhaps you need to sue OpenAI for libel. That might compel them to have someone lobotomize it on this topic to avoid a lawsuit.
So I went and asked chatgpt "is 2^2023-1 a prime number"
I did it multiple times. I expected either the correct "I cannot answer this question" answer, or pure comedy gold.
I got pure comedy gold. Not only half of the answers are "yes" and half "no", they include large glass of math lingo smoothie that sounds super impressive, is completely factually incorrect, and is presented in a very assertive tone.
Artifical conman... Is there such a thing as mind sickness? From consuming such output?
ChatGPT is the redditors of AI.
It will never answer "I don't know". It will only give you concrete answers it thinks is correct even though you know it is wrong.
I asked it again the same question this morning and the ChatGPT answer has changed somewhat - but still very wrong.
Prompt. what can you tell me about Canada Post's lawsuit against Geolytica inc?
ChatGPT. Canada Post filed a lawsuit against Geolytica Inc. in 2012 over allegations that the company's online postal code lookup tool, called "Postal CodeOM," violated Canada Post's copyright and trademarks.
Canada Post argued that the tool made unauthorized use of their copyrighted postal code database, which is protected under Canadian law. They also claimed that Geolytica's use of the Canada Post trademark in promoting the tool was misleading to consumers.
Geolytica, on the other hand, argued that postal codes are part of the public domain and that Canada Post had no legal right to claim ownership of them. They also claimed that their use of the Canada Post trademark was fair and legal.
The case went to trial in 2018, and in December of that year, the Federal Court of Canada ruled in favor of Canada Post. The court found that Geolytica had indeed infringed Canada Post's copyright and trademarks and ordered the company to pay damages of $500,000.
However, Geolytica has since appealed the decision, and as of my knowledge cutoff in September 2021, the appeal was still ongoing.
I mean what do you expect from bot which can't provide even correct answer to simple query like "illegal sites with direct download links", first it tells me sites with torrent links, then when I say torrent links are not direct download links but it should be file hosting links it provide me legal file hosting sites instead sites linking them, if I input such query to Google, already first two links are correct without any further clarification