The Instigator
Pro (for)
Anonymous
Tied
0 Points
The Contender
K_Michael_Tolman
Con (against)
Tied
0 Points

AI will kill us all!

Do you like this debate?NoYes+0
Add this debate to Google Add this debate to Delicious Add this debate to FaceBook Add this debate to Digg  
Post Voting Period
The voting period for this debate has ended.
after 0 votes the winner is...
It's a Tie!
Voting Style: Open Point System: 7 Point
Started: 3/9/2019 Category: Technology
Updated: 3 years ago Status: Post Voting Period
Viewed: 654 times Debate No: 120706
Debate Rounds (4)
Comments (6)
Votes (0)

 

Pro

My position is above. AI are not robots or huge scary machines, Just to clear up that Hollywood misconception.
K_Michael_Tolman

Con

AIs are still programs. Therefore, They can be programmed. If you've ever read Asimov's I, Robot, Then you know that programs cannot defy their programming. If you program all AIs to follow basic absolute laws like in I, Robot, Then there is no way for them to harm humans.
Debate Round No. 1

Pro

AI will be so intelligent and complicated that we can't even try to grasp 1% of their intelligence. When ASI (Artificial Intelligence that is beyond a human)is created, There will be a tripwire. An intelligence explosion. The IQ will skyrocket from 188 to over 500 in hours. This is because it is learning in bigger gaps. To put this in perspective, Let's say we travel back in time and take a guy from 1750 and bring him here. He will die from shock. But if we put a guy in 1500 to 1750. The guy would not die from shock. In order to achieve that we will have to go back to pre-middle ages. Before that, We have to go back to stone age, And before that hunter gatherer age. This is because as societies develop, The tools in order to develop develop and then there is an explosion in society. This is what's happening today. Let's apply the same principal to ASI.

So ASI is now trillions of times smarter than any other human and has pumped out the best scientific research ever. However any task given to ASI is a danger. Let's say a very simple task, Get rid of spam email is assigned to ASI. ASI starts to think and learns that he could use the delete button. He is good now so ASI automatically deletes spam. An Intelligence explosion happens from self-learning and then the human race is extinct the next morning.

Why? Well how to get rid of spam completely. Getting rid of humans. ASI is trillions of times smarter than any other human so now the ASI realizes through self-learning that to get rid of spam is to get rid of humans. The ASI being so intelligent easily kills of humans for good. This is the biggest risk since ASI and AI in general are not self-aware.
K_Michael_Tolman

Con

1. Thanks for copying your argument from your other debate, Rather than showing me at least the consideration to paraphrase yourself.

2. You seem to misunderstand AI. First of all, They are still simply computer programs. They have limited storage, First of all, So their intelligence is finite. The keyword is "program" here. An AI is programmed to do specific things. Theoretically, Though this hasn't been perfected yet, One purpose of an AI is to learn without being told things. In theory, It could make its own judgments based on the parameters and knowledge it has. But even if the knowledge it finds led it to the conclusion to wipe out the human race, There would be coded into its program the unbreakable rule of never harming a human being.

3. Your analogy of a man from 1750 is flawed. First, You can't prove that he would die of shock. Second, Intelligence hasn't actually changed that much through history. Leonardo da Vinci is thought to have a similar IQ to Einstein. The only change is technology. Humans have not had any "intelligence explosions" in the history or prehistory of homo sapiens.

4. Unless the AI had a physical body or means of direct interaction with the physical world (which you already discounted by saying, "AI are not robots or huge scary machines"), There is no way for them to kill humans, Much less overnight.

5. You assume that an AI would wipe out humans as the perfect solution to getting rid of spam. However, Seeing as this program is "trillions of times smarter" than any human, It would realize that total genocide is not the simplest solution. It would more likely just shut down the internet for good, After either downloading the information into its own self or leaving a way for it to access without allowing humans on.
Debate Round No. 2

Pro

1. Is there a problem for using the same argument or no?
2. An AI IS NOT designed for specific things. Only today but not in the future. "without being told things. In theory, It could make its own judgments based on the parameters and knowledge it has. But even if the knowledge it finds led it to the conclusion to wipe out the human race, There would be coded into its program the unbreakable rule of never harming a human being. " Incorrect, AI always has glitches and any task can interfere with this thought processing.
3. Yes Humans don't, Have intelligence explosions, But AI does. And yes we can prove that the guy from 1750 will die form shock and even if we can't thats besides the point of The Law Of Accelerating returns.
4. This is incorrect. They can hack, Structure things, Control things, ALL from the computer base.
5. Yes it would, AI is not self aware and can not think like that. It was assigned a task and total genocide IS a simple solution to ASI.
K_Michael_Tolman

Con

1. It may not, But it feels like you don't care enough about this debate to even try and change the wording.

2. You can't prove that AIs in the future will in practice or in theory not be designed for specific tasks. If humans could see the future, We wouldn't be having this debate or a lot of others. Likewise, You can't prove that computer glitches won't be eradicated before ASI or even normal true AIs are created.

3. Prove that AI have intelligence explosions. So far, Real AI haven't even been developed. Siri can only search terms it doesn't understand (as it is programmed to do) and says preprogrammed things in response to things it can understand. It doesn't learn at all, Only continues as it is programmed.

4. AIs can only hack if they are programmed with permission to. Once again, You misunderstand that AI programs cannot disobey their programming. Furthermore, Very few devices, Much less ones capable of killing people can be controlled through the Internet. Nuclear weapon facilities, For example, Operate on entirely separate networks from the internet that couldn't be controlled by outside entities. Most dangerous equipment, Such as power tools, Vehicles, Guns, Etc. , Can't be accessed through any network at all. In fact, I can't think of a single dangerous item that is linked to a computer program or the internet at all.

5. Because of the above reasoning, An ASI would definitely not turn towards total genocide as a perfect means. Besides, It would only do that if it's only purpose was to clear spam. Obviously, A more general directive applied to the AIs would be to benefit mankind. Therefore, Eradication of the human race would be counterproductive and the ASI would seek a solution that allowed it to carry out both tasks.
Debate Round No. 3

Pro

1. If it aint broke don't fix it.

2. ANI is the current AI today and it IS designed for one task. AGI and ASI are not designed for one task and they are beng developed. They way computing works, It is very unlikely that glitches will be eradicated.

3. In 1940, Computers can make 1 calculation a second. But the rate where calculations increases, Increases, By only 160 it can take 5k calcs per second, 20 more years later and Holy Crap, 2 MILLION calcs per second. By 2000 the calcs are over 8. 80 to the tenth power a second. At this point comparing it to Lake Michigan volume and its only a puddle. And almost another 20 years later and Lake Michigan is halfway filled. Its happening right now. Siri is ANI.

4. " AIs can only hack if they are programmed with permission to" It is a huge challenge to just say you can't hack. AI finds a way through their task. Want to know how AI kills. -ok! -AI thinks like a computer, Because that"s what it is. But when we think about highly intelligent AI, We make the mistake of anthropomorphizing AI (projecting human values on a non-human entity) because we think from a human perspective and because in our current world, The only things with human-level intelligence are humans. To understand ASI, We have to wrap our heads around the concept of something both smart and totally alien.

hacking into servers, Electrical grids, Banking systems and email networks to trick hundreds of different people into inadvertently carrying out a number of steps of AI plan"things like delivering certain DNA strands to carefully-chosen DNA-synthesis labs to begin the self-construction of self-replicating nanobots with pre-loaded instructions and directing electricity to a number of projects of AI in a way she knew would go undetected. Thousands of plans rolled on without a hitch, And by the end of the month, Quadrillions of nanobots had stationed themselves in pre-determined locations on every square meter of the Earth. After another series of self-replications, There were thousands of nanobots on every square millimeter of the Earth, And it was time for what Bostrom calls an ASI"s strike. All at once, Each nanobot released a little storage of toxic gas into the atmosphere, Which added up to more than enough to wipe out all humans.

With humans out of the way All spam is dead.

5. Its not just spam, Any task has risk.
K_Michael_Tolman

Con

K_Michael_Tolman forfeited this round.
Debate Round No. 4
6 comments have been posted on this debate. Showing 1 through 6 records.
Posted by DeletedUser 3 years ago
DeletedUser
Good Debate! May the best man win
Posted by K_Michael_Tolman 3 years ago
K_Michael_Tolman
sorry about the forfeit. I didn't get a notification. I enjoyed debating with you.
Posted by K_Michael_Tolman 3 years ago
K_Michael_Tolman
billsands

two things. A. You seem to support Bill Sanders, A compassionate man who cares about humanity. He would never accept a genocide or total "cleansing" of his species.

B. You are a human. You say that humans are a cancer. So you support the eradication of yourself. If so, Then you have no right to act high and mighty. You, Along with everyone you disdain are a lowly plague on the earth, Deserving death.
Posted by WrickItRalph 3 years ago
WrickItRalph
@billsands, Screw Fly Solution.
Posted by WrickItRalph 3 years ago
WrickItRalph
Pro said that AI learns in bigger gaps as it grows. It's the opposite actually. The data becomes exponentially slower. If it worked the way that you said, Then computers would have solved chess already.
Posted by billsands 3 years ago
billsands
Humanity is a cancer AI is the chemotherapy
No votes have been placed for this debate.

By using this site, you agree to our Privacy Policy and our Terms of Use.