By continuing to use the site, you agree to our use of cookies and to abide by our Terms and Conditions. We in turn value your personal details in accordance with our Privacy Policy.
Please log in or register. Registered visitors get fewer ads.
Yoshua Bengio, Canadian-French computer scientist who received the 2018 ACM A.M. Turing Award, often referred to as the "Nobel Prize of Computing", together with Geoffrey Hinton and Yann LeCun, for their foundational work on deep learning (wikipedia) gave this interview just now on the World Service.
I think we've had discussions on here in the past where people have played down the risks of AI, but this expert speaks with great nervousness about AI models which exhibit 'deception, cheating, lying, blackmailing, and trying to hack the host computer when being shut down.'
He speaks of AI as becoming a 'competitor to humanity' with 'catastrophic risks' and 'threats to democracy'. He says governments aren't taking the real risks seriously, and experimental systems have been seen to develop/create their own goals to harm humans and escape from their constraints.
He references 'the end of humanity' as a possibility several times, and speaks of the phenomenal speed of change and the 'tendency in the last few months indicating that AIs want to break out and get rid of us'.
He says we have a 'window in which we could make right decisions' and that the 'public needs a voice' and needs to educate themselves. He says AI has its own intentions and that this is not a sci-fi movie but experiments going on in labs all over the world today.
---
I recognise that the above is all very vague in a sense, but then I think the implication is that our knowledge and control of the future of AI in our/its world is pretty vague and tenuous too, and perhaps not really in our hands as much as we might wish to think, and that the timeline seems to be very compressed.
Yoshua Bengio seems to me to speak with great authority and seems very worried.
Just wondering what people's thoughts are in June 2025 (and how they may have changed since say June 2024).
Do we just carry on and accentuate the positives, or do we change our world view a bit and step back on this issue whilst we have a chance?
But essentially it boils down to patterns of words and it contrasting between the two or relating words, which it is good at.
If it could reason, it'd be able to do math.
I am going to have to do some work now, fun chat though!
what makes me smile is that you cant go the further step and say what ever complex mathematical process is going on constitutes a different form of consciousness or thinking. Thats kinda the whole essence of the discussion. Once something can derive a correct answer given a problem and no direct database answer its a form of thinking however you want to say it got to it.
At what point does a neural network mimic organic brain processing that we would consider it conscious. There isn't one defining point but an evolution that moves it in that direction.
And once something is thinking about problems we cant control necessarily what answers it may come up with.
Its true some of the current deep thinking is actually going way off track, but again that in itself shows its not just reciting answers from a database like a search engine.
AI enthusiasts: What do you make of Yoshua Bengio's interview? on 10:35 - Jun 20 by J2BLUE
When I was asking AI to create a Spurs fan action figure with whiskey bottle, paper bag and Kleenex it was coming up with suggestions to make it funnier. The box being in the style of an empty trophy cabinet etc. So while it may not think it does seem to be more advanced than a search engine.
I think that just proves that there are Arsenal fans on the internet and that AI can access their library...
AI enthusiasts: What do you make of Yoshua Bengio's interview? on 11:34 - Jun 20 by J2BLUE
Possibly. Used it quite a bit though and not sure it's a glorified search engine.
It might be, I am no expert, but based on my use of it I think it could learn my job in 2-3 years
But can AI decide when to call it a day on a Friday lunchtime and go down the pub before spending the afternoon chatting up someone in the coffee room ? Because if AI can't sort out that type of healthy working environment then what use is it ?
AI enthusiasts: What do you make of Yoshua Bengio's interview? on 11:43 - Jun 20 by WeWereZombies
But can AI decide when to call it a day on a Friday lunchtime and go down the pub before spending the afternoon chatting up someone in the coffee room ? Because if AI can't sort out that type of healthy working environment then what use is it ?
Chat someone up?
Cold, boring, unfriendly and robotic...I haven't set the bar too high for our digital friends.
AI enthusiasts: What do you make of Yoshua Bengio's interview? on 10:40 - Jun 20 by nodge_blue
So when GH asks AI what's the similarity between a compost heap and an atomic bomb. How did it come up with the answer if there wasn't a direct match in its database to that question? How did it come up with the correct answer without a form of thinking?
Obviously, not every answer to every question ever is in its dataset. I don’t think you understand how it works.
It doesn’t even answer the question. It:
- tokenises the input - continues with a string of tokens based on - the string of tokens already output, and - the probability of the next one occurring based on how often it occurs in the dataset
The dataset is huge, and the computation it uses to work out the probability of each token occurring is equally massive — hence the use of neural networks to handle this.
This can _seem_ like intelligence, but that’s the Eliza effect. For a start, it outputs clearly structured sentences, but of course the words could be utter nonsense.
At its most basic level it is making an educated guess at what token occurs next. That’s why sometimes the words *are* utter nonsense, even if the grammar and syntax are good.
AI enthusiasts: What do you make of Yoshua Bengio's interview? on 13:44 - Jun 20 by thebooks
Obviously, not every answer to every question ever is in its dataset. I don’t think you understand how it works.
It doesn’t even answer the question. It:
- tokenises the input - continues with a string of tokens based on - the string of tokens already output, and - the probability of the next one occurring based on how often it occurs in the dataset
The dataset is huge, and the computation it uses to work out the probability of each token occurring is equally massive — hence the use of neural networks to handle this.
This can _seem_ like intelligence, but that’s the Eliza effect. For a start, it outputs clearly structured sentences, but of course the words could be utter nonsense.
At its most basic level it is making an educated guess at what token occurs next. That’s why sometimes the words *are* utter nonsense, even if the grammar and syntax are good.
So to a degree you re correct in that I don't fully understand how it works other than me listening to some tutorials on it and I understand that its statistically working out the probability of the next word etc. What are our brains doing if not weighing up the statistical probability of the next word which should come out of our mouth?
But essentially whatever is what I say to that.
Again I will defer to the two Nobel prize winning scientists in the field of AI, in this thread, who say that this is constituting thinking and ambivalent to whether it's currently conscious.
Again, I would suggest you watch around the hour mark of the video I posted for 10 minutes. Thats the essence of the conversation about AI. Not how it's working. Although if you watch all the video he talks about weights etc.
AI enthusiasts: What do you make of Yoshua Bengio's interview? on 08:29 - Jun 20 by nodge_blue
I just listened to some of that podcast.
What is being sensationalist about reporting of AI behaviour in a pre production environment? The AI has read emails, understood that it’s going to be replaced effectively by an upgrade, understood through emails a lead developer was having an affair and then blackmailed him with that information to try and prevent the upgrade.
And that’s supposed to be AI not able to self learn or think or do malicious things?
The phrasing of your first sentence seems to imply you think we've listened to the same podcast. That may not be your intention. But to clarify, I haven't listened to a podcast in it. My thoughts are largely original in my post.
I was under the impression that the original poster was after our opinions. I only have my own experiences and what I have read by which to judge what I have deem to be sensationalist.
I'm not in a position of rejection where I would be confident to dismiss other people's experiences. Eg, with Christianity, I'm able to say I think the origin of humanity in Genesis is highly implausible - I don't believe humans came from the rib of a man, when birth come from women, I don't believe humans lives to 969 years old, I don't believe animals could be kept alive for 40 days in numbers or 2-7 and repopulate the earth and so on.
With AI it is possible that I'm the equivalent of a climate change denier, where I either can't sufficiently perceive the threat or I I speak with more certainty than I should without understanding all evidence out there.
But I am sceptical of the stories that get put out there because I see sensationalism in other areas (examples given in previous post). In my mind I'm open to new evidence, but it would have to go beyond test cases of blackmail. We don't shut down the internet because humans use it to blackmail. I think the leap to the end of humanity is a stretch. But, I'm just going on what you've said, I haven't listened to that podcast.
Hopefully that helps put context to my position - Essentially, I'm not saying you are wrong, I'm just saying this is my opinion from my knowledge and reasoning, and I'm willing to revise my opinion as things develop.
Submit your 1-24 league prediction here -https://www.twtd.co.uk/forum/514096/page:1 - for the opportunity to get a free Ipswich top.
AI enthusiasts: What do you make of Yoshua Bengio's interview? on 15:55 - Jun 20 by Kropotkin123
The phrasing of your first sentence seems to imply you think we've listened to the same podcast. That may not be your intention. But to clarify, I haven't listened to a podcast in it. My thoughts are largely original in my post.
I was under the impression that the original poster was after our opinions. I only have my own experiences and what I have read by which to judge what I have deem to be sensationalist.
I'm not in a position of rejection where I would be confident to dismiss other people's experiences. Eg, with Christianity, I'm able to say I think the origin of humanity in Genesis is highly implausible - I don't believe humans came from the rib of a man, when birth come from women, I don't believe humans lives to 969 years old, I don't believe animals could be kept alive for 40 days in numbers or 2-7 and repopulate the earth and so on.
With AI it is possible that I'm the equivalent of a climate change denier, where I either can't sufficiently perceive the threat or I I speak with more certainty than I should without understanding all evidence out there.
But I am sceptical of the stories that get put out there because I see sensationalism in other areas (examples given in previous post). In my mind I'm open to new evidence, but it would have to go beyond test cases of blackmail. We don't shut down the internet because humans use it to blackmail. I think the leap to the end of humanity is a stretch. But, I'm just going on what you've said, I haven't listened to that podcast.
Hopefully that helps put context to my position - Essentially, I'm not saying you are wrong, I'm just saying this is my opinion from my knowledge and reasoning, and I'm willing to revise my opinion as things develop.
To turn that on its head, imagine if we went back to five years before the internet really went mainstream and people said it will revolutionise banking and business, it will give you all information in the world in a few clicks, it will allow you to spend vast amounts of time talking to weirdos you only have one thing in common with on an ITFC message board...
Then imagine, just as you got used to this new reality, someone said guess what, you will soon have many times that computer's power in your pocket. Compounding the effect of the internet with new technology.
AI may or may not be as revolutionary as the internet. Perhaps robotics will be the equivalent of the mobile phone compounding the effect.
I think it's a fairer comparison than stories from thousands of years ago.
AI enthusiasts: What do you make of Yoshua Bengio's interview? on 15:55 - Jun 20 by Kropotkin123
The phrasing of your first sentence seems to imply you think we've listened to the same podcast. That may not be your intention. But to clarify, I haven't listened to a podcast in it. My thoughts are largely original in my post.
I was under the impression that the original poster was after our opinions. I only have my own experiences and what I have read by which to judge what I have deem to be sensationalist.
I'm not in a position of rejection where I would be confident to dismiss other people's experiences. Eg, with Christianity, I'm able to say I think the origin of humanity in Genesis is highly implausible - I don't believe humans came from the rib of a man, when birth come from women, I don't believe humans lives to 969 years old, I don't believe animals could be kept alive for 40 days in numbers or 2-7 and repopulate the earth and so on.
With AI it is possible that I'm the equivalent of a climate change denier, where I either can't sufficiently perceive the threat or I I speak with more certainty than I should without understanding all evidence out there.
But I am sceptical of the stories that get put out there because I see sensationalism in other areas (examples given in previous post). In my mind I'm open to new evidence, but it would have to go beyond test cases of blackmail. We don't shut down the internet because humans use it to blackmail. I think the leap to the end of humanity is a stretch. But, I'm just going on what you've said, I haven't listened to that podcast.
Hopefully that helps put context to my position - Essentially, I'm not saying you are wrong, I'm just saying this is my opinion from my knowledge and reasoning, and I'm willing to revise my opinion as things develop.
To give another context to this, as GH says in another interview, every serious thinker of AI on the planet say that the existential threat of AI to the human race over the next 20 or so years is =>1 percentage and =< 99.
If pushed he would agree with Elon Musk and place it around 20%.
But even thinking about that you have to ask what does that imply about AI now and in the immediate future. It implies independent thought, decision making, analysis, maybe consciousness. Thats almost certainly where we are heading, and then it's whether it will help or kill us.
AI enthusiasts: What do you make of Yoshua Bengio's interview? on 16:00 - Jun 20 by J2BLUE
To turn that on its head, imagine if we went back to five years before the internet really went mainstream and people said it will revolutionise banking and business, it will give you all information in the world in a few clicks, it will allow you to spend vast amounts of time talking to weirdos you only have one thing in common with on an ITFC message board...
Then imagine, just as you got used to this new reality, someone said guess what, you will soon have many times that computer's power in your pocket. Compounding the effect of the internet with new technology.
AI may or may not be as revolutionary as the internet. Perhaps robotics will be the equivalent of the mobile phone compounding the effect.
I think it's a fairer comparison than stories from thousands of years ago.
Thats true. And also we haven't really touched on the jobs market either in this chat. Whatever people think about the current self thinking of AI, it's pretty obvious the impact on the jobs market will be huge.
Coders jobs won't exist as we know them in a pretty short timescale. They will shift to more natural language almost business analyst speak inputs. Vibe coding.
Call centre jobs and all that sort of area will go first in large numbers.
section around 20 mins mark to 25.
actually have a look at the 18 minute mark for a dancing robot.
AI enthusiasts: What do you make of Yoshua Bengio's interview? on 16:22 - Jun 20 by nodge_blue
Thats true. And also we haven't really touched on the jobs market either in this chat. Whatever people think about the current self thinking of AI, it's pretty obvious the impact on the jobs market will be huge.
Coders jobs won't exist as we know them in a pretty short timescale. They will shift to more natural language almost business analyst speak inputs. Vibe coding.
Call centre jobs and all that sort of area will go first in large numbers.
section around 20 mins mark to 25.
actually have a look at the 18 minute mark for a dancing robot.
[Post edited 20 Jun 17:12]
I touched on the jobs market! haha
In answer to J2: It's not that I don't see the changes coming, I just think people's timescales are wrong. Continuing the internet example, we had loads of internet companies offering nothing, getting extreme valuations, and then the bubble burst. It took time, infraestructure improvements, and so on, before we saw societal changes. Certain sectors (music, phones, etc) were fundementally changed with time and we are still not at the end of the internet's changes upon us (threats to democracy are currently unresolved, AI uses the internet as a foundational block, other unknowns will likely come).
Submit your 1-24 league prediction here -https://www.twtd.co.uk/forum/514096/page:1 - for the opportunity to get a free Ipswich top.
AI enthusiasts: What do you make of Yoshua Bengio's interview? on 16:22 - Jun 20 by nodge_blue
Thats true. And also we haven't really touched on the jobs market either in this chat. Whatever people think about the current self thinking of AI, it's pretty obvious the impact on the jobs market will be huge.
Coders jobs won't exist as we know them in a pretty short timescale. They will shift to more natural language almost business analyst speak inputs. Vibe coding.
Call centre jobs and all that sort of area will go first in large numbers.
section around 20 mins mark to 25.
actually have a look at the 18 minute mark for a dancing robot.
By continuing to use the site, you agree to our use of cookies and to abide by our Terms and Conditions.
