4 ms·

I get the impression this story is warped quite a bit by the source Jason Rohrer. If I understand correctly, the so called "censorship" were safety precautions

I get the impression this story is warped quite a bit by the source Jason Rohrer. If I understand correctly, the so called "censorship" were safety precautions for AI development. The response of some posters seems to be to discard the safety topic alltogether as something "old", that "neanderthals" do. I think it is exactly those "curious" people AI research has to worry about most, who do not take the time to think about the consequences and possible outcomes of what they are doing. Maybe a chatbot will not gain superpowers and take over the world, but it can surely do lots of other harm. Then, when AI tech moves beyond narrow AI and includes more abilities, this will in fact lead to desaster, as the same careless people with the greatest "curiosity" will throw caution overboard as was the case with the narrow AI before... It's not a slippery slope, we are right in the territory AI safety researchers warn about for a decade now, write lengthy books about and try to create awareness for in the public and governments.

option_greek 5y ago
Humans barely agree with each other. Do you think are going to be persuaded by half baked chat bots like GPT3? (ya the cherry picked stuff looks good but rest is still not really there). So far all OpenAI has done is generate more publicity (probably intentionally may be with an eye towards investors) by acting as if they are protecting the next nuclear weapon of some kind. Frankly, there is nothing to lose by not using their "ohh god its so dangerous" model.
- NicoJuicy 5y ago
  The difference is the scale. You are not convicting people they reply on you. But for everyone that response there could be 100's of readers.
- mdoms 5y ago
  Your premise is that people are not suggestible? I think the past decade of social media has conclusively origen that wrong.
- jbuhbjlnjbn 5y ago
  Machine learning is synonymous with non-linear growth. So of course I think any chatbot will eventually have no problem persuading humans. The pivetal moment is when his persuation skills succeed that of the average human. I have no clue how far away we are to that goal, but given accelerated growth, it will come rather sooner then later.
  - option_greek 5y ago
    We are already being convinced by invisible bots that have no mouths to talk with and no keyboards to type with. Go to any website -> all the ads that are desperately trying to woo us or nothing but AI/machine learning. Go to Amazon -> the product placement and results customized to you -> AI/machine learning. Considering Google/Amazon are already making boat load of money selling stuff indirectly and directly through this, I would say no new danger is coming by the way of Zeus's thunderbolt (aka GPT-3 as per ClosedAI's attitude). Unless you are a bot and managed to convince me to spend a couple of minutes typing this answer - in which case - well played bot, well played! :D
- Zababa 5y ago
  > Humans barely agree with each other. Do you think are going to be persuaded by half baked chat bots like GPT3? Isn't that precisely why we can be conviced by things like GPT3? Humans don't value other humans opinion that much, so it's easy for people to trust bot over other humans.
- SpicyLemonZest 5y ago
  I think any chatbot good enough that a guy can "turn it into a close proxy of his dead fiancee" is capable of persuading people, yes.
schleck8 5y ago
i bet they asked to integrate a service flagging insensitive outputs, which is a valid request in my opinion, given not only their company image is on the line but also that of text to text transformers in general
snowwrestler 5y ago
The harm at risk is not just to humans but to chat bots and AI projects. Remember Tay from Microsoft? It lasted about 24 hours before it started to return offensive replies. MS got embarrassed, had to take it down, and that was that. Unsupervised learning systems are more efficient and powerful today. Are they any better at navigating the many cultural sensitivities and restrictions in human society? “We Regret to Inform You GPT-3 Is A Nazi” is a headline that would probably collect a lot of clicks, and instill (or reinforce) a public perception that such technologies are worthless or dangerous. Does GPT-3 have the internal safeguards necessary to protect itself from being abused in this way? I doubt it. I suspect this is one reason OpenAI asks that people using GPT-3 supply such safeguards themselves.

option_greek 5y ago

Humans barely agree with each other. Do you think are going to be persuaded by half baked chat bots like GPT3? (ya the cherry picked stuff looks good but rest is still not really there). So far all OpenAI has done is generate more publicity (probably intentionally may be with an eye towards investors) by acting as if they are protecting the next nuclear weapon of some kind. Frankly, there is nothing to lose by not using their "ohh god its so dangerous" model.

NicoJuicy 5y ago
The difference is the scale. You are not convicting people they reply on you. But for everyone that response there could be 100's of readers.
mdoms 5y ago
Your premise is that people are not suggestible? I think the past decade of social media has conclusively origen that wrong.
jbuhbjlnjbn 5y ago
Machine learning is synonymous with non-linear growth. So of course I think any chatbot will eventually have no problem persuading humans. The pivetal moment is when his persuation skills succeed that of the average human. I have no clue how far away we are to that goal, but given accelerated growth, it will come rather sooner then later.
- option_greek 5y ago
  We are already being convinced by invisible bots that have no mouths to talk with and no keyboards to type with. Go to any website -> all the ads that are desperately trying to woo us or nothing but AI/machine learning. Go to Amazon -> the product placement and results customized to you -> AI/machine learning. Considering Google/Amazon are already making boat load of money selling stuff indirectly and directly through this, I would say no new danger is coming by the way of Zeus's thunderbolt (aka GPT-3 as per ClosedAI's attitude). Unless you are a bot and managed to convince me to spend a couple of minutes typing this answer - in which case - well played bot, well played! :D
Zababa 5y ago
> Humans barely agree with each other. Do you think are going to be persuaded by half baked chat bots like GPT3? Isn't that precisely why we can be conviced by things like GPT3? Humans don't value other humans opinion that much, so it's easy for people to trust bot over other humans.
SpicyLemonZest 5y ago
I think any chatbot good enough that a guy can "turn it into a close proxy of his dead fiancee" is capable of persuading people, yes.

schleck8 5y ago

i bet they asked to integrate a service flagging insensitive outputs, which is a valid request in my opinion, given not only their company image is on the line but also that of text to text transformers in general

snowwrestler 5y ago

The harm at risk is not just to humans but to chat bots and AI projects. Remember Tay from Microsoft? It lasted about 24 hours before it started to return offensive replies. MS got embarrassed, had to take it down, and that was that. Unsupervised learning systems are more efficient and powerful today. Are they any better at navigating the many cultural sensitivities and restrictions in human society? “We Regret to Inform You GPT-3 Is A Nazi” is a headline that would probably collect a lot of clicks, and instill (or reinforce) a public perception that such technologies are worthless or dangerous. Does GPT-3 have the internal safeguards necessary to protect itself from being abused in this way? I doubt it. I suspect this is one reason OpenAI asks that people using GPT-3 supply such safeguards themselves.