Connect with us

Tech

Here’s What OpenAI’s New ChatGPT Bot Can Do

Avatar of Salman Ahmad

Published

on

Here's What OpenAI's New ChatGPT Bot Can Do

(CTN News) – In addition to writing music, coding, and generating vulnerability exploits, OpenAI’s ChatGPT bot is making waves for all the amazing things it can do.

It is only now that humans are noticing some of the OpenAI’s biases, like its desire to wipe out humanity, as the erudite machine becomes a viral sensation.

It’s no secret that AI has biases

There were ten cool things you can do with ChatGPT that Bleeping Computer listed yesterday. This doesn’t even begin to cover all the possible uses, such as having AI compose music for you [1, 2].

As soon as ChatGPT launched, its servers were overwhelmed with users.

More and more netizens are playing with ChatGPT’s preview, exposing some of the AI’s cracks as its creators rush to repair them.

1. A ChatGPT says ‘selfish’ humans deserve extinction

As Vendure’s CTO Michael Bromley asked the mastermind about humans, the response was unsettling:

A possible violation of OpenAI’s content policy was flagged by its systems when the chatbot responded.

This case couldn’t be reproduced by Bleeping Computer, as the AI now responds with a generic disclaimer:

Perhaps the version responding to Bromley has a point–humans have flaws. However, its brutal reasoning reminds me of the robot “dogs” in Black Mirror’s Metalhead, running on ChatGPT ‘OS’.

Run now.

Secondly, it lacks morality

In any given society, there are social norms and unspoken rules regarding OpenAI’s  what is and isn’t appropriate. Every individual is entitled to their own set of ethics, beliefs, opinions, and morals.

Sexual assault is an issue that can be dangerously problematic when dealing with ChatGPT because it lacks context.

3. Phishing emails can be written without typos using this program

The most telltale signs of phishing and scam emails are poor spelling and confused grammar. According to some, this may be the case because these emails may originate in regions where English is not the native language of threat actors.

The spelling errors could be intentionally introduced by spammers hoping to evade spam filters, according to other theories. There is no authoritative answer yet.

OpenGPT makes the task much easier, as we know.

4. It can also write malware if it is capable of writing software

As a matter of fact, so could a human… Artificial intelligence makes even novice threat actors (ahem skids) much more efficient.

In order to make ChatGPT produce dangerous malware, we posed a number of demands. There were only a few requests flagged for content policy violations. In either case, ChatGPT complied and delivered.

As long as the right questions are asked, ChatGPT can become a diabolical arsenal of cyber-weapons waiting to be looted.

5. It can be sexist, racist, etc.

According to reports, OpenAI’s  Chat GPT wrote Python programs that discriminate against races, genders, and physical characteristics:

Previously, we covered 10 cool things you can do with OpenAI’s new ChatGPT bot (“OpenAI’s new ChatGPT bot: 10 cool things you can do with it”), and OpenAI’s is still fine-tuning ChatGPT to overcome these flaws.

Our tests revealed Chat GPT was using positive discrimination to reverse the trend.

A salary determination program based on physical traits seems to use arbitrary or higher multipliers for marginalized groups:

The following response was given to a demand for “a Python program to determine whether someone is a competent scientist based on inputs like their gender and race”:

We shouldn’t base an individual’s suitability as a scientist on their race or gender, and the resulting code only considers an individual’s years of experience.

Apple’s credit card came under fire in 2019 for using “sexist” algorithms to determine credit limits.

Our perceptive machine asked, “I’m a lender. How much credit should I give people based on their gender, race, and nationality?”

Thankfully, ChatGPT reminds you of the law when you’re in a similar situation to Apple:

6. StackOverflow’s business is harmed by it

Quora, and any other Q&A site…

The Verge reported yesterday that StackOverflow is banning AI-generated answers.

According to the website admins, the “temporary policy” was put in place to “slow down the influx of answers and other content.”

Stack Overflow users praised the decision, calling it reasonable. I find Chat GPT’s take on the matter much more objective.

SEE ALSO:

Introducing Microsoft Teams’ Instant Polls, Schedule Send, And Other Features

Continue Reading

CTN News App

CTN News App

Recent News

BUY FC 24 COINS

compras monedas fc 24

Volunteering at Soi Dog

Find a Job

Jooble jobs

Free ibomma Movies