Tech ethics group wants investigation into OpenAI's GPT training

Hanno Labuschagne

Journalist
Staff member
Joined
Sep 2, 2019
Messages
5,339
Tech ethics group wants investigation into OpenAI's GPT training

A prominent tech ethics group plans to file a complaint with the US Federal Trade Commission urging the regulator to halt further commercial deployment of new generations of artificial intelligence technology that powers the popular OpenAI Inc. tool ChatGPT.

The complaint from the Centre for Artificial Intelligence and Digital Policy, which is led by longtime privacy advocate Marc Rotenberg, centres on GPT-4, the language model behind the research group Open AI’s ChatGPT, a computer program designed to convincingly simulate human conversation.

[Bloomberg]
 
As someone who dabbles in a bit of machine learning from time to time, albeit for much simpler tasks than something like ChatGPT, it's amazing how ignorant and naive some people in tech, who should know better, sound like when talking about ChatGPT, or artifical intelligence in general. ChatGPT is nothing more than a very sophisticated machine learning model. Like any machine learning model, it's only as good as the data it is trained on. It's not sentient, it's not going to take over the nukes, and it certainly isn't going to go rogue on us.
 
As someone who dabbles in a bit of machine learning from time to time, albeit for much simpler tasks than something like ChatGPT, it's amazing how ignorant and naive some people in tech, who should know better, sound like when talking about ChatGPT, or artifical intelligence in general. ChatGPT is nothing more than a very sophisticated machine learning model. Like any machine learning model, it's only as good as the data it is trained on. It's not sentient, it's not going to take over the nukes, and it certainly isn't going to go rogue on us.

"it's only as good as the data it is trained on"
And that's the point.
I work in the space where generative AI is highly prevalent.
And the quality is utter shite.
It's just keyword salad, trying to fool Google's algorithm.
Facts? Who needs facts? Key points just get made up, or the prevailing falsehood is not corrected.

Let this carry on, and soon the majority of the net will be fake info.
 
Back
Top