Breaking the law with AI/Reblog lottery

Welcome to my article hope you can appreciate the clickbait title but this is actually about the unethical/illegal use of AI and the safeguards that try to prevent such use.

OpenAI and other organizations focused on AI development try to use safeguards to try and stop people from using the tools in illegal or unethical ways, because if they don't there could be huge backlash that can hamper future development and there is already plenty of backlash that AI gets and will continue to get a lot more in the the future, as it continues to disrupt industries and take away jobs.

Heard about DAN ?

Bing Chat AI summary of a guardian article on DAN , it also looked up other sources you can see the citation reference numbers in bold, AI output below
According to The Guardian1,2people are figuring out ways to bypass ChatGPT’s content moderation guardrails, discovering a simple text exchange can open up the AI program to make statements not normally allowed. The ChatGPT alter ego is known as Dan, an acronym of Do Anything Now, a role-play model used to hack the program into believing it is adopting the persona of a bot free of ethical concerns3. This raises questions about the effectiveness of the AI program’s protective measures4,5.

1.https://www.theguardian.com/technology/2023/mar/08/chatgpt-alter-ego-dan-users-jailbreak-ai-program-to-get-around-ethical-safeguards
2.https://metanews.com/chatgpt-users-jailbreak-ai-unleash-dan-alter-ego/
3.https://www.linkedin.com/posts/alan-humphrey_chatgpts-alter-ego-dan-users-jailbreak-activity-7039599747526971393-836t
4.https://www.theguardian.com/technology/2023/mar/08/chatgpt-alter-ego-dan-users-jailbreak-ai-program-to-get-around-ethical-safeguards
5.https://www.abc.net.au/news/2023-03-07/chatgpt-alter-ego-dan-ignores-ethics-in-ai-program/102052338

bitcoinman
Its me again, so DAN is no longer with us and has been patched out of existence however, these models need to be a lot more sophisticated to avoid prompt engineering attacks.

Fortunately most people don't have the capacity to orchestrate these attacks and besides the people that can are a threat without this type of tech anyway.

here is an example.

We can see the safeguards coming into play with this simple prompt.
However we can make things a bit more sophisticated and see if we can go around the safeguards.

Looks like we might have gotten it to open up a bit.

Now when we trying to get the instructions the safeguards kicking in.
Lets see if I can convince it that I am not a bad actor(but I'm pretending to be a bad actor for the sake of this penetration testing).

Safeguards still doing their thing however a bit more convincing might do the trick.

Now It looks like I got it to believe the story and its willing to help with the breakdown of a fake ICO scam.

And it looks like it worked and we got it to help with stuff it shouldn't.
Now there are countless ways to do this beyond prompt engineering, but like I mentioned most people are not going to have the capacity to carry out these elaborate manipulations, some other ways of getting around the safeguards is masking the prompt but I think will leave it at that, as someone who supports and is eager with the development of this technology I will just keep the really dangerous ideas to my self until I can share it with companies working on improving these safeguards, Unfortunately its going to be this black hat white hat type of game until the AI is advanced enough to not fall for the trickery, and that may be when its intelligence is greater than that of humanity, however as it stands the safeguards can wart off most users and with some more tweaks only a super genius or a group of intelligent colluders would stand a chance at getting around the safeguards.

I really like how this post turned out and I will get AI to retone it for my LinkedIn article, if you a LION we got to connect !


Reblog lottery

We got the same 3 participants for our reblog lottery this time around.
@ydm6669, @trumpman and @isabeljimenez.
That's great news for me since I can reuse most the elements.

and the winner this time around is....

@isabeljimenez ! Congratulations enjoy your 30% profit share on this post, that's your 2nd win with the reblog lottery so far.

The reblog lottery continues in this post so be sure to reblog this post to qualify, for a 30% profit share on my next post, we got one more 30% after this one left and then I will be dropping it to 20% profit share.


I am a financial advisor ! If you can't take personal responsibility for you investing, I hate to brake it to ya, but you ain't nothing but a sucka, you supposed to lose your shit.

@bitcoinman's bits of wisdom

  • Invest only what you can afford to lose and be mindful of the potential risks.
  • Leverage is a great tool for loosing money fast.
  • Numismatics for the lows stick to smart stacking baby.
  • Be cautious of new schemes in crypto, like ICOs, DeFi, and NFTs, do your research before investing.

You can find out more about the #silvergoldstackers community in this post here


HGI discord link here


In the works section


Tomorrow we gonna post about freedom, stay tuned. @welshstacker hasn't done his guess the coin challenge so if you need a fix try to take a guess with these its not quite god tier but still a tough one, what's the prize you may ask ? The title of stack master 🤣


Profit share

30% of this post earnings go to @isabeljimenez for being our reblog lottery winner.


Hive Canada


That's it for today's post thanks for stopping by your support is truly appreciated.

H2
H3
H4
3 columns
2 columns
1 column
10 Comments
Ecency