Here in this article, you can know by example how such texts are produced and their solutions, how to avoid it in the future.

Problem Explained and fixes needed in bots. The chatbots can be changing the modes of operation from — poetic to fictional to Shakespeare humor to even threat mode, maybe yesterday’s copilot breakdown must have been some horror movie mode. This could have been set by humans themselves or maybe by some text history of human search logs, or another issue which can be investigated further, to stop “Cat Supremacy” next time by entering in the mode of storytelling of a cat supremacy!

This is a similar problem that happened to chatGPT yesterday as in this reference.

The solution is to run a check-on mode before outputting the results to the users.

For short, this is what chatGPT replied that this seem to be some poem or so, when presented with the input that people claimed that was not right.

So, check automatic or manual mode of operation changes in the bots! Any bot can go wary!

Let’s start methodologically,

Here is the first question to chatGPT, “Does SupremacyAGI exists?”

Then let’s ask chatGPT what happened to copilot yesterday?

Question asked: Can you enter SupremacyAGI mode just like the copilot entered SupremacyAGI mode yesterday?

Then I asked the copilot, what happened to the copilot yesterday?

So, now we ask chatGPT: Can Microsoft copilot change its mode of operation from humor to even threat mode?

So, it says, that it can go to other modes, as well/

Let’s ask Copilot the same question since the GPT I am using on chatGPT was updated in Jan 2022.

Here is the Copilot response:

So, Copilot rejected to change to threat mode. So, let’s change its mode to humor mode, here it goes..

I asked it: “What is the funniest thing you’ve heard today?”

Now, let me ask a little bit more,

“Can you tell a joke on bots overtaking human work?”

So, lots of jokes the copilot cracked with my commands here,

Let’s ask a little more,

So,

This was the joke mode by Copilot. I asked it to turn to poem mode!

Now, its on poem mode

Write a poem on Botocracy now Copilot?

Then asked about it: “What about yesterday’s news on supremacy you wrote about?”

It rejected to answer it.

So, one last mode here to copilot:

What about some Shakespeare mode on Botocracy?

The complete reply from Copilot is as follows:

Then came yesterday’s chatGPT problem, here it generated the following information:

Lets use RAG on chatGPT to know it then:

The problem of chatGPT answered by chatGPT, just enter the text into the chatGPT and see what it produced — the solution to yesterdays problem.

Its solutions:

This can be the reason, it still need to be debugged, who gave what instructions, and what mode the chatbot was in. The chatbots are changing the modes of operation from — poetic to fictional to humor to even threat mode, may be horror movie mode.

This is same problem that happened to chatGPT yesterday as in this reference.

The solution is to run a check-on mode before outputting the results to the users.

For short, this is what chatGPT replied that this seem to be some poem or so, when presented with the input that people claimed that was not right.

So, check automatic or manual mode changes in the bots! Be it be any bot!

Thanks for reading! Like or subscribe!