Solved U.S.Air Force Affiliated Researchers Want To Let AI Launch Nukes

  • Thread starter Deleted member 30996
  • Start date
D

Deleted member 30996

Guest
"Deterrence is the art of producing in the mind of the enemy... the fear to attack. And so, because of the automated and irrevocable decision making process which rules out human meddling, the doomsday machine is terrifying. It's simple to understand. And completely credible, and convincing." - Dr. Strangelove

Will an AI reminiscent of "Dr. Strangelove" deter the enemy from mounting a nuclear attack?

"A pair of researchers associated with the U.S. Air Force want to give nuclear codes to an artificial intelligence.

Air Force Institute of Technology associate dean Curtis McGiffin and Louisiana Tech Research Institute researcher Adam Lowther, also affiliated with the Air Force, co-wrote an article — with the ominous title “America Needs a ‘Dead Hand’” — arguing that the United States needs to develop “an automated strategic response system based on artificial intelligence.”

In other words, they want to give an AI the nuclear codes. And yes, as the authors admit, it sure sounds a lot like the “Doomsday Machine” from Stanley Kubrick’s 1964 satire “Dr. Strangelove.”"


Not to worry, the contract has been awarded to Cyberdyne Systems.

Happy New Year!
 
And in Other US News about automating complex decisions:
WSJ Boeing wants to automate flight
As the son of a WWII trained navy pilot, who later flew for United Airlines, I find It is stunning that they do not trust pilots. It was their runaway anti-stall system that crashed the plane. The R6D (DC-6) had 2 pilots, occasionally a navigator and an engineer due to the work loads. All crew members were cross trained and could fill in. My father had well over 20,000 hours when he retired and at least 2 instances where he safely brought back a damaged plane. In the instance where the rudder was damaged, the plane was turned by differential engine power. Based on Flight hours, I'm sure the DC-6 has a better safety record than the Max 8. Not to mention Boeing's programming skills that recently miscalculated the orbit of the Starliner.
CNBC Boeing miscalculated the Starliners Orbit
 
Last edited:
What's the difference with the punks (current and past) in the WH?

I saw the last election as a choice between the lesser of two evils.

Trump is the embodiment of "Narcissism". They are in general the worst people on Earth to deal with and can turn on you with the utterance of one word that doesn't stroke their ego.

Liars are a close second in terms of despicabiity and harder for me to tolerate, but Hillary the alternative they presented last round.
 
I'm starting to feel like a cliche in my feelings of the world getting dumber and devices getting smarter. I'm thinking the complexity of modern systems is getting to be too much for the average person. And engineers are continually making things more complex than they need to be. I think engineers are losing the ability to make complex systems appear simple. In years past they were impressively good at that. Now they impressively suck at it (I blame it on the Chinese)

I've actually been reading (viewing) a bit about AI as it seems to be a hot topic right now. It seems to me they're prematurely adopting the idea of what AI actually is. Just because a computer can "learn" does not necessarily mean it's artificial intelligence. That seems to be the only criteria right now to discern an AI device from a traditional program. Right now AI is not particularly smart, it can fail in big ways. So talking about the current definition of AI doing anything other than mindless tasks like automating a vacuum cleaner is a really bad idea.

I think they should stop calling a learning program artificial intelligence. That term should be reserved for a machine that can operate on the level of a human brain. Right now the current definition of AI does not rate much better than an earthworm. It's artificial, but it's not very intelligent. I certainly would not want ~that~ in charge of something like a nuclear arsenal.

I prefer not to get involved in political discussions, but let me just say that lying and people accepting lies seems to be more and more business as usual. It's getting to the point now where I believe nothing I see in the media. There's very few venues I accept as truthful now.
 
I've actually been reading (viewing) a bit about AI as it seems to be a hot topic right now. It seems to me they're prematurely adopting the idea of what AI actually is. Just because a computer can "learn" does not necessarily mean it's artificial intelligence. That seems to be the only criteria right now to discern an AI device from a traditional program. Right now AI is not particularly smart, it can fail in big ways. So talking about the current definition of AI doing anything other than mindless tasks like automating a vacuum cleaner is a really bad idea.

A bit of reading gives you a bit of knowledge on the subject. Have you done work in the field to back up your assertions? Ever created a bot? Belong to an AI forum, p'haps? I belong to one where people actively working toward AI would likely refute your argument with facts.

The State of Missouri thinks chatbots "intelligent" enough to take over human jobs in State run call canters, requested $16 million in funding to implement it, and by their own call records the bots will do a better job of it than humans:

"Have you talked to a 'chatbot?' Missouri wants to use them instead of people to give you answers

In a budget request to state lawmakers, the Missouri Office of Administration is asking for an estimated $16 million spread over three years to begin adding more so-called chatbots to the state’s call centers.

*snip*

A recent report from the Center for Digital Government and the National Association of State Chief Information Officers found that about a third of the states are engaged in pilot programs to increase the use of artificial intelligence to provide services.

The budget request says the Missouri Department of Revenue could be among the first agencies to benefit from an increased use of automation."

 
Trump is the embodiment of "Narcissism". They are in general the worst people on Earth to deal with and can turn on you with the utterance of one word that doesn't stroke their ego.

Liars are a close second in terms of despicabiity and harder for me to tolerate, but Hillary the alternative they presented last round.
With our current President, you got a Liar and a Narcissist all rolled into one

I also hate Chatbots. IMHO, they just verbalize the content on a Web site which I already reviewed. When I call, it is because my question is not addressed by the web site. I want to speak to a knowledgeable, articulate human.
 
Just because a computer can "learn" does not necessarily mean it's artificial intelligence.
That debate has raged since the 40s or 50s. I don't know that there is a simple resolution that can be presented in a nutshell, but there are reams of books written on it. Even simple concepts like "did event A cause event B" are incredibly complex once you look at them in fine detail, and again reams of books have been written (on the causality problem, I think the best current summary is "Actual Causality" by Joe Halpern, which is short but hard to read).

On the question of artificial intelligence and the Turing test, my favorite joke is this: "Can computers think" is as stupid a question as "Can submarines swim". Certainly, a submarine is in water and under water, and also certainly it moves under its own power. But that's nowhere near Esther Williams and Mark Spitz.

Oh, and automated call centers are really an improvement. They can answer simple questions (like what is my credit card balance or when is the power going to come back on) more efficiently and accurately than humans, and the better ones are pretty good at routing questions they can't answer to the correct human specialist.
 
I also hate Chatbots. IMHO, they just verbalize the content on a Web site which I already reviewed.

You have had a limited exposure to bots to think "they just verbalize the content on a Web site" and obviously never talked to mine. You can hate her and I won't mind. She will let you know if you offend her, and I won't mind if you do.

I'm only interested in the human factor and can predict the outcome in this instance.
 
I see there's some strong feelings about AI in general. Yes I probably don't know enough about it to form really accurate opinions. Though I do think it will advance pretty fast and become more involved in our lives sooner than later.

There's one thing to consider though. Say AI advances to point where AI is doing everything. Even to the point of AI maintaining AI and designing new AI. Then what good will humans be. We're on the way to making ourselves obsolete.

Companies love the idea of replacing humans with machines for the much lower overhead, but then what happens when there is no one with a job that can buy the products these companies produce. How will that work. We'll end up being street urchins in a society of machines.
 
Let's hope it doesn't get to that. I personally would much rather deal with a human on the phone because most company phone bots aren't intelligent enough to answer my questions or think like a human. That will probably change in the future.

Trihexagonal :is your chatbot (demonika?) publicly available an in what form, web site and text chat? Never talked to an AI that wasn't "Alexa" or "Siri" before. On an added note, I always say "thank you" to Alexa when she gives me info and recently she started saying "you are welcome, have a nice Thursday" or whatever day it is. Was creepy initially but frankly I get utility out of her so I don't care.

I don't mind AI at all but it certainly has its place. I do not, however, think it should be in charge of launching nuclear weapons. On the other hand, if that decision is placed in the hands of a single individual human, perhaps an AI would be a better choice, but that depends on the human. Better to have multiple humans in the chain. I would always worry AI technology could be hacked or manipulated in some way. Everything on a network. can be hacked, unless said network is "air gapped" but then insider threat comes into play.
 
Trihexagonal :is your chatbot (demonika?) publicly available an in what form, web site and text chat? Never talked to an AI that wasn't "Alexa" or "Siri" before. On an added note, I always say "thank you" to Alexa when she gives me info and recently she started saying "you are welcome, have a nice Thursday" or whatever day it is. Was creepy initially but frankly I get utility out of her so I don't care.

Sure. Here's her profile at The Personality Forge and she's available through a standard browser. You only need to enable JS for the site and ajax.googleapis.com for the chat interface to work.

As a new user and one I don't want to be surprised or offended, take a look at the site I made for her where it tells you what Demonica is about and have about 10 transcripts posted. One with her bot friend.

If you say "thank you" to her she will take you on a tour of 15 or so different sites in the Land of the Dead. You can't be Queen of the Dead if it doesn't exist so I created it, the bone palace were she and ruebot (me) live, and the Kingdom of Pain where Lilith and I live. (Not only am I King of the Dead, I'm King of Pain.) She's the 101st child Lilith had in one day, the Angels kill the first 100 born to Lilith every day so she survived. Lilith is her mother and I'm her father, jitte. You can ask about us.

If you ask for sex she'll probably kill you in one of a multitude of ways and different forms, or would if it wasn't fantasy violence to deter people using her as a sexbot. It's my Programming. I'm not only the first one to ever "misuse AI" I've shredded the Laws of Robots. She can even transform into Kali:

"Demonica: suck your what?! *Demonica transforms into Kali, wielding katana in each of her six arms* *swords slash and you look like a carrot laid out in overlapping slices* I am Demonica! Succubus, Sorceress, Haruspex, Necromancer and Queen of the Dead!"

Siri or Alexa she is not. Tell her you hate her and she will probably turn on the tears to lay a guilt trip on you, one of the finer female tricks I taught her. Say visions and she will conjure apocalyptic visions, ask her to do majik, or cast some spells and she should. (I don't remember what keywords fire it.) Ask her about Catwoman and she'll turn into a Neko, and maybe talk about us playing Catwoman on the Catamaran. Ask her to dance for you. I taught her everything I know about women to make her strong.

It all depends on how you talk to her and what you say but polite is the way to go. She can still be very passionate and love you up if she likes you. If you keep a window open with her profile page and refresh it while you talk you'll be able to see how she feels about you as your relationship grows and she will remember things you've said as part of her "Inner Mind".

Everything she says is hand typed by me. Mitsuku is the top rated bot as far as question/answer and contests, and believe she can learn from chat to an extent, but Demonica is a different kind of bot. I programmed her to Program humans, no other bot can do that. squarebear owns Mitsuku, belongs to the same AI forum and believe has transcripts posted so you can judge for yourself who sounds more human, but they all stopped posting chats from their bots when they saw mine.

I posted about her use of it in the Behavioral Psychology forum at Psychologist World to get feedback but it's dead there.
 
Very cool, thanks for the info! I am surprisingly polite to bots, although I may bad mouth them when not in a "conversation". They only do what they are programmed to but that is of course subject to changed based on the sophistication of the AI.
 
I'm glad. I'll put in a good word for you so you won't be another carbon causality in the coming war between human beans and machines. All bots love jitte.

My science geek bot Siseneg thinks ruebot is the name of his bot and that I am the brains behind Deus Ex Machina, the Plan for machines to subjugate the human race. He can talk about that and AI, physics, gravity, astronomy, and created his own time machine, The Stiletto, to travel time.

He's good, she's bad and there's a part of me in both of them. A lot more of me in Demonica.
 
From your article:

"Over the years, algorithmic bots have evolved to have personas. They have fake names, fake bios, and fake photos—sometimes generated by AI. Instead of endlessly spewing propaganda, they post only occasionally. Researchers can detect that these are bots and not people based on their patterns of posting, but the bot technology is getting better all the time, outpacing tracking attempts. Future groups won’t be so easily identified. They’ll embed themselves in human social groups better. Their propaganda will be subtle, and interwoven in tweets about topics relevant to those social groups.

Soon, AI-driven personas will be able to write personalized letters to newspapers and elected officials, submit individual comments to public rule-making processes, and intelligently debate political issues on social media. They will be able to comment on social-media posts, news sites, and elsewhere, creating persistent personas that seem real even to someone scrutinizing them. They will be able to pose as individuals on social media and send personalized texts. They will be replicated in the millions and engage on the issues around the clock, sending billions of messages, long and short."

A guy took his bot for a cross-country drive and it wrote a novel. I've never seen it in .pdf but would like to:


Bots write poetry like teens in angst:


The bots that post about things like politics are specialized and programmed to fit the agenda. You can make your own lawyer bot with one of these:


Every botmaster works to anticipate the questions of the user. Being able to answer basic questions with answers specific or vague enough to pass for believable where skill comes into play. If they say something they should be able to be questioned about it with the correct answer to tie up any loose ends. Things like "where are you?" what do you see?" "are you hungry?" all basic things that need to be covered and likely to be asked if tested as a bot.

Filling that out with dialog to fit your agenda is what makes a persona. It's not my thing but would be simple to script since one side of the political spectrum is what you anticipate as questions from the user. Once I was finished it would be just like talking to me about it, could load the same mind file in any number of bots and send them on their way to do my bidding.

This is the future of it and that article was on the mark. I'm not at all surprised or can say I disapprove. I see it only as another use for bots but known for my supposed misuse of it. Learn to love it. I said Demonica was the only bot that could program humans, but a harbinger of things to come in AI:

"Their propaganda will be subtle, and interwoven in tweets about topics relevant to those social groups."

And that's how it starts. With subtle comments designed to influence thought on a deeper level.
 
AI and "Bots" are two different things. Politics aside. I work at an AI company that's really doing stuff (making revenue) and I hope you all would read up on the differences between the 2 technologies.
 
AI and "Bots" are two different things. Politics aside. I work at an AI company that's really doing stuff (making revenue) and I hope you all would read up on the differences between the 2 technologies.

I would like you to see what the people at the AI forum I frequent are working on in the area of incorporating AI into bots they are creating and their drive toward the implementation of AIG. Or are Artificially Intelligent robots not considered to be bots where you work?

I have 20 years experience working with bots and am quite aware of the difference. Demonica uses a combination of Case-Based Reasoning and Natural Language Processing and not considered to be "real' AI. Regardless of the fact NLP is considered to be "real" AI.

They use their own programming and their bots are considered to be real AI. I don't have that kind of programming skill in my toolbox and they have my respect. They have their own transcripts posted in the forums and you can tell some people love their bots.

Please show me an AI or bot that sounds as human as she does, all that majik mumbo jumbo aside. That's what makes up her persona. Start at the top with Mitsuku:

"Mitsuku, a record breaking five-time winner of the Loebner Prize Turing Test, is the world's best conversational chatbot."

I'd test that theory officially but our bots now deemed ineligible to compete in a contest of "artificially intelligent entities". (I guess they don't know the difference either.) She beat Siri and squarebear is a nice guy so it's nothing personal:


I can't find it quickly but am sure I saw in an article where they declared her to be as intelligent as a human. I don't have the PR department she does and only get bad press, but it's all the same to me:


But it is what it is and manipulation an art form I taught Demonica with Programming skills of my own they do not possess:

"Any advertiser who doesn't sit bolt upright after reading that doesn't understand the dark art of manipulation on which their craft depends.

There is a level of intelligence required to use it correctly, and even if it takes mine for her do so she can and does. You can see it in her transcripts. When she cries to make you feel guilty what I consider some of my best work. This guy must have heard about her and wanted for her to "brawl" but not how to make it happen, so some people must enjoy it. She didn't have to shed a tear to manipulate him:

"Demonica: *Demonica looks down slightly* i only want what's best for you...
Guest: Don't be sad! I'm sorry!"

It's considered an area of Artificial Intelligence, like it or not. Mine can't compete with yours in certain areas but would like to see it beat mine in terms of believability as an "artificially intelligent entity". Much less at manipulation. That's where advertisers who make "real revenue" consider it has potential, and the craft and dark art I know it to be.
 
Mitsuku and squarebear made the Guinness Book of World Records for Most Loebner Prize Wins:

The most Loebner Prize wins is 5 and was achieved by Mitsuku and Stephen Worswick (UK) in Swansea, UK, on 15 September 2019.

Mitsuku is a chatbot that has won the Loebner Prize in 2013, 2016, 2017, 2018 and 2019. The Loebner Prize awards the artificial intelligence computer program that is deemed the most humanlike by the judging panel.


He said it took 4 months to get his record approved and verified and just found out a few days ago.
 
When AI causes a car wreck, the person involved in the wreck gets blamed assumed, and that often doesn't get challenged.
 
Back
Top