Neural Net, Large Language Model (LLM) and related technologies that power ChatGPT, Google’s Bard AI and other generative AI platforms are rooted in programs that anyone of sufficient technical know-how and computing resources can make use of.

And as generative AI is showing its power and ability to grow user bases and generate profits, others are looking to spin up AI systems that accomplish their own designs and goals.

Two recent initiatives have recently made news. One involves “TruthGPT,” a proposal by Elon Musk to build a generative AI platform that seeks to expand knowledge and quest for “truth” without political or other biases.

The other involves something called ChaosGPT, the spin-up of a generative AI system which has been coded with objectives of wiping out humanity.

TruthGPT: A Quest For Unbiased Truth, Or Filling Market Demand For Less Woke AI?  

Elon Musk has said “TruthGPT” would be designed to be “maximum truth-seeking.”

He has criticized systems such as OpenAI’s ChatGPT, for being “trained to lie” and for being “closed source” and “for-profit.”

Musk says his intentions with TruthGPT would be to build a more ethical and beneficial AI system, based on open source protocols and programming, and focused on understanding the nature of the universe without biased guardrails. He has said that he hopes to create a third major option to OpenAI and Google in the AI space, and that he believes TruthGPT could “create more good than harm.”

The world leading tech entrepreneur was one of the original co-founders of OpenAI in 2015, the company that created ChatGPT.

But Musk left the project in 2018.

Musk has said he was concerned and wanted OpenAI to be more aggressive in its research and development, and that he was frustrated with the company’s decision to pursue a for-profit business model.

Musk also had disagreements with Sam Altman, who now leads the company. Musk believed Altman was not being transparent enough with him about OpenAI’s plans.

Musk also expressed concerns that AI could pose a serious threat to humanity if it was not developed responsibly. Musk has said he was not convinced OpenAI was doing enough to address those concerns.

In any case, OpenAI, especially given its alliance with Microsoft, and Google are far along in the race for advanced generative AI.

Whether TruthGPT will be able to emerge and successfully compete with these first movers remains an open question.

Forget Helping Humanity: ChaosGPT Is Tasked With Wiping It Out 

For those who believe any form of “strong AI” will end up posing serious dangers for humanity, the frank goal of the clandestine programmers (or “prompt engineers”) behind ChaosGPT to wipe out humanity is actually oddly refreshing.

The actors behind the dark initiative utilized Auto-GPT, an open-source AI project that allows anyone to use natural language prompts to set up tasks or goals for the generative AI system.

As shown (at least for awhile) in a YouTube video, a AI persona called “Chaos-GPT” was given five tasks:

  • to destroy humanity
  • establish global dominance
  • cause chaos and destruction
  • manipulate humanity
  • attain immortality

Thus “booted up,” the AI began attempting to carry out the assigned objectives.

As and others reported, ChaosGPT churned away, gathering information from the internet, including performing searches and saves related to its objectives.

It attempted to recruit other AI to help with its mission, including soliciting another AI to study lethal weapons. That AI said it wanted peace and would not assist. 

ChaosGPT returned to Googling, in pursuit of information regarding its quest.

The experiment drew attention and plenty of criticism. More than 10 thousand followed a ChaosGPT account on Twitter before it was taken down. 

AI Can Be Spun Up By Nearly Anyone

Auto-GPT, an open source project accessible via GitHub, describes itself as “An experimental open-source attempt to make GPT-4 fully autonomous.”

It automates chatbot use. Utilizing the software, users can assign a role to an AI, establish five objectives, and begin. Auto-GPT will then develop its own prompts for ChatGPT, utilizing it to first build a strategy and then execute it via the leading generative AI created by OpenAI and now partnered with Microsoft.

The project allows the combining of ChatGPT, its own capabilities, and live internet data and software installation.

But Auto-GPT is just one option available to those who wish to experiment with their own generative deep-learning neural net AI systems.

Before ChatGPT burst into public awareness in November 2022, The Trends Journal covered the public release of software which gave almost anyone with some decent computing power the ability to spin up an AI. (See “NOW ANYONE CAN DOWNLOAD AGI THAT CAN OUTHINK HUMANS,” 23 Aug 2022.)

Companies like Databricks offer the abilities for organizations to implement in-house generative AI systems using LLM (Large Language Model) and other technologies.

[Full disclosure: details in this article involving Elon Musk’s prior involvement with OpenAI were gleaned by querying a “neutral” third party…you guessed it: Bard AI. —JD]

TRENDPOST: Expect an explosion of competing generative AI projects.

There will also be growing regulatory frameworks. But those frameworks may not be substantially concerned with preventing unbound scientists from trying to introduce self-aware, autonomous AI which can outperform humans in every respect–so-called “strong AI” or “AI Singularity.”

Instead impositions and guidelines so far have focused on instilling politically correct conformances in AI.

These restrictions, exemplified in the Biden Administration’s “AI Bill of Rights,” are designed to substantially control the “thought content” that AI produces, and to act as yet another gatekeeping limit on human political expression and freedom. (See “BIDEN ADMINISTRATION SUBVERTS CONSTITUTION WITH ‘AI BILL OF RIGHTS’,” 11 Oct 2022.)

Skip to content