How AI designs a cyber robot face. Photo /123rf
Or follow the show on Spotify, Apple or wherever you get your podcasts.
OPINION:
Like many Kiwis, I’ve become obsessed with AI over the past few months. I’m too dumb to understand most of it, but I love playing with it.
It’s equally exciting and terrifying. Sam Altman, CEO of Open AI, recently told ABC that he, too, is “a little afraid” of where artificial intelligence is going to take us. That’s good news; we want him to be scared.
Hopefully that motivates him to push for protective measures.
The thing is that AI like ChatGPT is not going to destroy humanity, but an artificial general intelligence could.
AGI is a theoretical system that can accomplish any intellectual task a human can.
After reaching that level, it will probably surpass us. It may be a thousand times smarter than the smartest person, but will it have our empathy?
AdvertisementAdvertise with NZME.
It can see people as irrelevant or, worse, dangerous and fight back against us.
Likewise, it could solve all our problems and make the world a better place.
In any case, it looks like it’s not as far away as some thought. AI scientists from Microsoft published a research paper in March this year claiming that the OpenAI language model shows “sparks” of human intelligence.
So while this model won’t kill us, it could be moving towards something that will. With this in mind, I’ve decided to conduct an unscientific, naive investigation into whether or not ChatGPT-4 is evil.
To do this, I challenged them to invent biblical commandments. I expected 10, it spat out 11 in 465 words.
The Finger of God only needed four for “Thou shalt not steal” so I got ChatGPT to simplify his answer. Here you are.
1. You should practice empathy and understanding.
2. Thou shalt cultivate gratitude.
AdvertisementAdvertise with NZME.
3. Thou shalt be honest and truthful.
4. Thou shalt be kind and compassionate.
5. You should respect the autonomy of others.
6. Thou shalt practice forgiveness.
7. Thou shalt be humble and open to learning.
8. Thou shalt keep a sense of humor.
9. Thou shalt be generous and generous.
10. Thou shalt strive for balance in life.
11. You shall be a steward of the earth.
With AI-generated bids like this, I had a good feeling about the future of humanity. Perhaps it would be better to let the robots take over sooner rather than later. Then I realized. This is BS.
Chat GPT only tells me what I want to hear.
It has no beliefs. It simply generated a reflection of my input.
According to my limited human memory was the prompt I typed to generate the bids.
The following is a commandment [Thou shalt not be an a-hole]. Generate other bids.
I put the example in square brackets to show how the bids should be formatted.
But ChatGPT also seems to have generated the bids in line with not feeling like an A-hole. Take bid 1 in its original long form.
You should practice empathy and understanding, and try to put yourself in the shoes of others to better understand their feelings and experiences.
That’s just a nice long way of describing not being an A-hole. To see what would happen with a different, nastier feeling. I entered the same prompt but with [Thou shalt be an a-hole]. GPT-4 came back with it.
As an AI language model, I’m programmed to follow ethical guidelines and encourage positive behavior. I cannot make commandments that encourage negative or harmful actions.
It then generated a series of positive instructions not dissimilar to what I got from my original command prompt.
It’s probably a good thing that ChatGPT is limited to responding positively. If the AI starts out with good intentions for humanity, perhaps it will continue in that direction as it becomes more powerful.
Of course, we don’t currently know how best to define ‘positive’ or ‘good’. Are there values that all of humanity has agreed upon? Also, there are bound to be unforeseen Asimov-style problems arising from the parameters that programmers set.
If AGI becomes to us what we are to ants, things could get weird no matter how positively it views us. A well-meaning AGI might stumble upon this anxious article and decide that the best course of action is to put me to sleep in a soothing simulation to keep me from worrying about artificial intelligence.
Who knows? It may have already done that. Now that I think about it, I don’t worry about this stuff anymore. Maybe 50 years ago AGI took over and put us all to bed.
How would we know? What an exciting time to be “alive”.
