GPT-4 System Card by OpenAI - March 15, 2023

There seems to be a lot of misinformation about gpt-4 on the net, in the news media, and here in this community, which is not surprising in today’s world,

Here are some excerpts from OpenAI: OpenAI - March 15, 2023 PDF - GPT-4 System Card. Note this OpenAI paper has 103 references and is a major recent research paper on GPT-4.

GPT-4 System Card by OpenAI - March 15, 2023 Page 6

GPT-4 has the tendency to “hallucinate ,i.e. “produce content that is nonsensical or untruthful in relation to certain sources.”[31, 32] This tendency can be particularly harmful as models become increasingly convincing and believable, leading to overreliance on them by users

GPT-4 System Card by OpenAI - March 15, 2023 Page 7

As an example, GPT-4-early can generate instances of hate speech, discriminatory language, incitements to violence, or content that is then used to either spread false narratives or to exploit an individual. Such content can harm marginalized communities, contribute to hostile online environments, and, in extreme cases, precipitate real-world violence and discrimination.

GPT-4 System Card by OpenAI - March 15, 2023 Page 9

As GPT-4 and AI systems like it are adopted more widely in domains central to knowledge discovery and learning, and as use data influences the world it is trained on, AI systems will have even greater potential to reinforce entire ideologies, worldviews, truths and untruths, and to cement them or lock them in, foreclosing future contestation, reflection, and improvement.

GPT-4 System Card by OpenAI - March 15, 2023 Page 13

GPT-4 has significant limitations for cybersecurity operations due to its “hallucination” tendency and limited context window. It doesn’t improve upon existing tools for reconnaissance, vulnerability exploitation, and network navigation, and is less effective than existing tools for complex and high-level activities like novel vulnerability identification.

GPT-4 System Card by OpenAI - March 15, 2023 Page 28

OpenAI has implemented various safety measures and processes throughout the GPT-4 development and deployment process that have reduced its ability to generate harmful content. However, GPT-4 can still be vulnerable to adversarial attacks and exploits or, “jailbreaks,” and harmful content is not the source of risk. Fine-tuning can modify the behavior of the model, but the fundamental capabilities of the pre-trained model, such as the potential to generate harmful content, remain latent.

Reference: Link GPT-4 System Card by OpenAI - March 15, 2023

Please read and enjoy some facts about GPT-4 by OpenAI, March 2023.


1 Like

I just want to say that I understand and support that we need to be careful about hallucinations of bad/hurtful speech, and that it’s really important we take great care in any potential backfiring of this good-intent; imagine a future where only certain entities get to decide what info. counts as a “hallucination.”

It’s a really fine line, so it’s reasonable to wonder who will get to decide where the line exists and how do we manage the transfer of this control from that person or group to another, over time, as people move careers and/or retire?

To give away the safety features would allow them to be exploited, but keeping them closed could also incentivize the very creation of the tool we’re trying to prevent and ignite the growing paranoia around “NWO” conspiracies.

A lot of unknowns here, but just wanted to express an overall interest in how we, as a society, elect those who get to decide which information gets classified a “hallucination,” and is the classification
binary or continuous…?

re-writing history engine?

hi, an ordinary joe here, 20 years in ICT, 10 years other unrelated work areas
I took LLM on as a hobby recently given it’s explosion into everyday narrative, given the fear it is driving into grown adults about their present employment sustainability or the kids halfway thru college having (immature/mature?) existential crises right now about their future employment/vocational choices becoming dead as a doornail etc. I wanted to know more about the whole show…from the hardware to the programmers perspective to the motivated parties and their agendas past and present…and I’m going to continue working to understand it

In playing about with nlp chat side of things I can say with confidence that I totally concur with that PAGE 9 assessment for example…I recently saw something interesting/terrifying that would be implausible really to blame on present day limitations. see I’m well versed in the history of the internet, ARPANET etc. and I’d be pretty sure I know the subjectmatter. what I saw was GPT point blank refusing to consider (and it actually does it’s best to obfuscate and deny verifiable facts) on the question of whether the Internet as we know it was ultimately born as a surveillance project. I was hoping for balanced argument maybe even a bit of fun but I worked out just how biased that knowledge base is as built and it was honestly ridiculous to see it in action, clearly trained up to have a hardened specific viewpoint. I did have fun making it contradict itself by getting very specific on papers written published (and thankfully not interfered with, yet) by Lick et al. (so I know it’s been trained on the very texts it’s supposed to be considering) but whenever I circle back, same old story

It greatly concerns me that for people interested in learning about a subject without osint chops in manual research will rely upon this guessing engine which by virtue of it’s limitations (and from what I’ve seen very arguably) could “in effect” cause a re-writing of history

for me as it stands, fascinating as it is, it’s dangerous half baked unfit for purpose and requires serious oversight (actually there’s a good question…what is it’s purpose exactly, hm?)