ChatGPT

[madscijr]

ChatGPT ... carries the potential of knowing EVERYTHING human kind has discovered since our existence.  

But does it? What about all the documents, videos/pictures/recordings not digitized? What about all the things that happen / have happened in the world that people have not written down in digital form and put on the Internet?

Most critically, what if ChatGPT is "trained" with false or subjective information, or information specifically written or edited to further an agenda? What about information supplied to it by racists, or a political party or government or organized crime or hackers or simply ignorant people?

It would be very easy to lie to it, or feed it certain information, to influence its knowledge base. It's easy to "pollute" these things.

Consider the Chinese or Russian governments, which stridently and effectively censor the information that hundreds of millions of people have access to. Or even the USA or any country where the media is controlled by certain interests. Consider the influence hackers had on elections by posting to Facebook.

Even in the free world where access to information is not controlled, people are easily influenced. How would you prevent that being done to a dumb AI?

One of the first ones you and I will likely be faced with is Education. Do we need to send our kids to school?

I don't know if you have kids of your own, but one of the most important parts of kids going to school (or any activity) is socialization. If you keep your children at home being spoonfed all their knowledge of the world by some computer program (which can be manipulated by hackers or other bad actors) will they really be getting a quality education and be adequately socialized?

And what about the quality of life? Up until now, kids go to school, make friends, get to know their teachers, participate in activities (including physical ones where they get exercise and learn to coordinate their bodies), and make memories in life. Are you saying that sitting at home all day talking to a computer chatbot can adequately replace that? I cringe to think of a world like that.

And if we have them wear VR goggles, so they can experience being with other people, there are some real concerns there:

First, anything computerized and on a network is open to manipulation. How do you know your kid is seeing what you think they're seeing?

Second, it's open to so many privacy issues. In VR, every single move you make can be tracked. If you participate in a virtual reality environment, not only can what you say, hear, type or do onscreen be tracked, your body movements can. Whatever your eyes look at, whatever body movements you make, they'll know. That might be acceptable to a totalitarian government like China or for tracking sex offenders, but do we really want that to be the norm for everyone?

That kind of information can be analyzed and used to manipulate someone - by a hacker, a predator, a government, or criminal. Do we really want to make people that vulnerable and subject them to that kind of invasive surveillance as the norm?

But mainly, this approach would effectively remove kids from reality, and limit them to a very controlled and limited field of experience, both inadequate and dangerous.

With ChatGPT could they not get all the knowledge they need to be whatever they want to be.

We've already talked about how this "knowledge" can be suspect, but beyond that, as it is, we already have a ton of knowedge and information on line at our disposal, but people are still misinformed and ignorant.

Even the most critical thinker can be misled by bad information when it's too widespread -or deliberately manipulated. And the world is loaded with that, even before computers.

What about people who don't make the effort to fully research their information, or don't have the time? As it is, the information out there in the mainstream news every day is slanted to direct people's attention both to get viewers and readers for ad revenue, and to forward the agendas of whatever parent corporation ultimately owns that media source. And people either don't know to, don't care to, or have the time to, verify the information with due diligence.

What is going to protect people from that when it's multiplied by the power of these AI systems?

Anyway, those are some thoughts I've been having as ChatGPT has arrived and become popular. I do think these kind of tools have the potential to be of immense value to humankind (and life on Earth in general), but we need to tread lightly, and foresee the dangers before we fully jump in. And we probably should never FULLY jump in, and replace good old reality with these things anyway!