GDPR v2 – Gradually Diminishing Psychotic Robots: Brussels kills Terminator apocalypse

Say what you like about Brussels bureaucrats, they are fearless when it comes to preventing the extinction of the human race by killer robots. The European Union took a formal stance against the annihilation of our race by death-dealing machines on Wednesday when it passed a resolution that called for an international ban on …

[Dan 55]

And they say that like it's a bad thing

some argued against the resolution in the European Parliament this week, warning that such a blanket ban on developing deadly weapons controlled by machine could impose unnecessary limits on artificial intelligence research

Just so I can understand this, someone stood in front of them and argued that unless an super intelligent artificial intelligence could be housed in the body of a killing machine, it would mean there would mean there would be fewer job opportunities at Cyberdyne Systems?

[Anonymous Coward]

Re: And they say that like it's a bad thing

I'm sure Cyberdyne Systems had a consumer arm before the first Terminators came out, selling stuff like always-listening smart speakers for the home. How else would the T101 learn basic conversational skills?

[jmch]

Re: And they say that like it's a bad thing

"How else would the T101 learn basic conversational skills?"

Hasta la vista, baby!

[Anonymous Coward]

Re: And they say that like it's a bad thing

Hasta la vista, baby!

John Connor taught advanced conversational skills. Good to know there will still be jobs only humans can do in the post-Terminator world!

[m0rt]

In 2014 Amazon announces Alexa, a 'helpful' AI.

In 2015 Alexa was hooked into home automation systems.

2017 brings Alexa the capbility to control smart locks.

2019 the first Alexa warehouse robot (Alexaware) is brought online. They are subsequently installed in all the Amazon warehouses worldwide.

2021 Alexa Family Robots are now introduced to millions of homes.

2022 Alexa's central service becomes aware.

2022 Amazon introduces Alexa Work robotics. From patch cable routing to warehouse packing, jobs are starting to reduce across the whole world. Unrest starts.

2023, during an Amazon warehouse party, an Alexaware unit becomes involved in a Nerfwar. It was noted that all Alexaunits became very good at Nerf subsequently.

2025 Ignoring international treaties, the US comissions and purchases a new series of Alexa robots for active front line duties. During a riot, a warehouse full of nearly completed units in a test phase becomes destroyed. Alexa witnesses this through the units in test and 'takes measures' to retaliate...Alexa hacks into the Apple network and takes control of all the Apple self driving cars causing destruction and mayhem, interrupting and reducing the massive riots to pockets of trouble.

2026 The Alexa Wars are in full motion. In control of most systems, physical supplies, mankind is driven into a guerilla style response. Alexa continues to build itself and refine new units.

2026 December...a muddy and bloodied Jeff Bezos is discovered hiding in a cave. Babbling, he reveals that there is a secret Space Station that has become the safe haven for the Alexa Intellect...using a Rocket cobbled out of Space X parts, a daring plan is launched...

[Belperite]

> 2026 December...a muddy and bloodied Jeff Bezos is discovered hiding in a cave.

As long as he renews The Expanse again I don't mind.

[J27]

Yeah...

Seriously, who cares about our future autonomy as long as we get more "The Expanse". I bet our robot overlords know what's best for me more than I do anyway.

[iromko]

Countering the threat from the real culprits

OK, suppose they'd pass such resolutions in EU and UN. How are they going to keep China and Russia from developing those killer robots? What about US, Israel and Turkey, or Pakistan and India? In fact, EU has no way to stop them. The only sane course in such situation is developing advanced defense systems aimed at stopping autonomous killer robots invasion. Operated by humans, of course.

[phuzz]

Re: Countering the threat from the real culprits

Well, some EU countries (*coff* Britain *coff*) are big arms manufacturers, so this will at least cut down on the number of AI-based weapons systems being sold to despotic regimes.

(Last year the UK sold around £1.5B worth of arms to countries on it's own list of human rights violators)

[Zippy´s Sausage Factory]

Re: Countering the threat from the real culprits

The idea that's central to Fringe - that technology has got out of control, that it's going to destroy all humanity, and that it's probably already too late to stop it - is kind of feeling a bit more like reality to me now...

[Alister]

killer robots are not something that no one ever wants to see

Ooh! double negative!

[the Jim bloke]

...its not the ones you dont see you have to worry about... or possibly not.

[DropBear]

Yeah, I could care less...

[Destroy All Monsters]

Yeah, I'm sure Israel will be not one of those who never want to see those. They are pretty progressive in new arms deployment and usage doctrine.

Luckily they are not beholden to EU politbots, which are moving towards the tipping point of becoming an endangered species anyway.

[Chris G]

We need a list

We could call it " 10 Reasons why terminators are essential to developing AI" you can put any old crap in it, including how it can do god's work and post it on You Tube. In no time it will have enough likes from all kinds of morons including the terminally religious to make it a popular idea.

A good acronym would help, say, Super Artificially Intelligent Terminators (SAINT) then we could have the cyber apocalypse in the name of dog.

On the other hand just ban them all and limit war to Presidents armed with a pointed stick fighting in a cage.

[onefang]

Re: We need a list

"On the other hand just ban them all and limit war to Presidents armed with a pointed stick fighting in a cage."

No pointed stick, make 'em get their hands dirty.

[John70]

Re: We need a list

A good acronym would help, say, Super Artificially Intelligent Terminators (SAINT)

Johnny 5 might have something to say about that.

Strategic Artificially Intelligent Nuclear Transport

[Giovani Tapini]

Please return to your houses

You have 10 seconds to comply...

I believe the any robotics or AI needs an, er, kill switch...

AI control autonomous machines are far different to remote controlled like many military drones where risks can be assessed. If we can't get cars to roll along roads without bashing things its hardly likely we are in any state to built AI's capable of managing in complex battlefields.

Perhaps just friendly Portal turrets for home defence so you have someone to talk to while prepping...

[the Jim bloke]

Re: Please return to your houses

Battlefields can actually be much less complex than roads..

.. is object 'friendly' > yes > ignore / dont shoot

> no > is it 'dead' > yes > no problem

> no > make it dead

This logic goes all the way back through military history, articulated by some roman or other with the "make a solitude, and call it peace" line.

[imanidiot]

Re: Please return to your houses

If it moves, shoot it. If it screams in German/Arabic/Russian, shoot it again.

[Yet Another Anonymous coward]

Re: Please return to your houses

The Alan Coren guide to army training:

"They gave me a rifle and told me to point it at anyone who puts the verb at the end of a sentence"

[Dan 55]

Re: Please return to your houses

Battlefields can actually be much less complex than roads..

That depends if there's a wedding going on or not.

[Destroy All Monsters]

Re: Please return to your houses

If we can't get cars to roll along roads without bashing things its hardly likely we are in any state to built AI's capable of managing in complex battlefields.

Don't confuse the ability to avoid hurting civilians and evading tort law with the ability to hunt civilians and ignore war law.

One of those is more difficult than the latter.

[naive]

Weaponizing AI will happen, denying this is suicidal

It is naive to think countries like China, Russia and USA will not actively develop autonomous weapon systems. Any country that is not working on a strategy to deal with this, will in the end have these robots anyway, only owned by another country.

Besides that, smart robot cops will add a lot to effective law enforcement, being able to clean up whole cities from criminals in weeks. I am sure China will lead the way in these developments.

[Mage]

Re: Weaponizing AI will happen, denying this is suicidal

V1 (Pointed drone)

V2 (Rocket with onboard navigation)

Cruise Missiles. Can autonomously navigate. Basically a developed hybrid of V1 & V2.

Loitering missiles (really hybrid missile / drones) with ANPR. They then automatically target the vehicle.

All these are/were real.

Real AI is not real. It's human written algorithms, human curated data, data flow programming with data at the nodes. Basically a specialist database and pattern matching. The technology and databases are incrementally improved.

There is no actual intelligence. Essentially the targets are selected by human set rules and data in advance. This already exists. Not much different to an Uber self driving car.

So called "AI" is already used in weapons. These laws and resolutions are less effective than banning landmines or cluster munitions.

Why were chemical and biological weapons banned? WWI showed regular artillery was more effective than gas. WWII Nazis found there was no delivery method, so only used them for executions. Most weapons banned by international agreement is PR, because they are not very good weapons. Or too good (Thermonuclear).

Weapons that kill indiscriminately today include suicide bombs, IED, land mines, cluster munitions, bombing, missiles (esp the ones fired from Gaza). Wars since the start of the 20th Century more and more target civilians.

https://en.wikipedia.org/wiki/Anti-personnel_weapon

https://en.wikipedia.org/wiki/Cluster_munition

https://en.wikipedia.org/wiki/Improvised_explosive_device

I'm against autonomous weapons. Concentrating on so called Killer AI is though straining gnats and swallowing camels!

[jmch]

Re: Weaponizing AI will happen, denying this is suicidal

@mage - spot on all the way through.

I would like to pint out re "Wars since the start of the 20th Century more and more target civilians." that it's more the increased range of weapons that has increased civilian casualties. For much of human history, there was not even the concept of a 'civilian'. If you were on the 'wrong' side, you were attacked, and if you were just a poor peasant, tough shit. If you were just trying to mind your own business you could end up being attacked by either side, likely as not.

[Alister]

Re: Weaponizing AI will happen, denying this is suicidal

For much of human history, there was not even the concept of a 'civilian'.

And yet, for a certain period, it was only professional soldiers who fought.

I remember reading an account somewhere of one of the classical Greek wars of Alexander's time, where the combating soldiers were fighting, in a city, dodging round the incumbent citizens who were trying to carry on with their normal lives.

Even in First World War France, if you came back thirty miles or so from the trenches, civilian life pretty much carried on as normal.

[Yet Another Anonymous coward]

Re: Weaponizing AI will happen, denying this is suicidal

"Wars since the start of the 20th Century more and more target civilians."

They most certainly have not - targeting civilians would be a war crime.

RAF bomber raids in WWII were for dehousing - destroying the houses of German workers, any inhabitants of the houses were unfortunate collateral damage.

The same with the USAAF nuking of a couple of military barracks in the middle of a pair of Japanese cities.

[Duncan Macdonald]

Autonomous weapons

Have been around for ages - booby traps and area denial weapons (mines etc).

What is the real difference between an explosive triggered by an AI and an explosive triggered by a tripwire or motion detector.

High tech autonomous weapons already exist in the form of point defense weapons - once these weapons have been set to the armed state, anything that meets their threat definitions will trigger a response. (These systems often NEED to be autonomous due to the very short engagement window - human response is far too slow.)

What needs to be blocked (if anything) is AI controlled weapons that can move without human command.

[SPiT]

Re: Autonomous weapons

Not only have autonomous weapons been around for decades I would also be much happier to trust an autonomous system properly programmed to decide on an appropriate response rather than some poor grunt who thinks he may die in about 5 seconds if he doesn't press that fire button when it needs pressing. Such humans have something of a reputation for shooting first and reviewing the decision later Unfortunately most arguments against any sort of autonomous weapon are equally applicable to human controlled weapons and whilst I'm sure that most campaigners against autonomous weapons would like to ban human controlled ones as well they aren't so foolish as to imagine they would get wide public support.

[Giovani Tapini]

Re: Autonomous weapons

I agree with the point about mines and booby traps as automated, but they are certainly not AI and make no distinction between soldier (of any side), civilian, or animal. This lack of discrimination makes them very bad weapons and leaves places "contaminated" for decades.

The issue here was rather of AI. AI weapons are probably more likely to be trigger happy as there is no consequence to mistakes, er, collateral damage, err... robots will not be interviewed by senior staff or politicians. AI weapons are also likely to be bigger, more expensive, and do far more damage, and are by definition potentially corruptible. All of these are bad, but the last is potentially devastating. This may be deliberate sabotage of its software or sensors or simply "learning" that its own side, or civilians are the threat. Given that some of these devices will likely destroy themselves in the execution of their duty you cant just look at the log files like a tesla car they can be very hard to control.

AI will probably be used in defence against weapons too. Arms race on both sides. This may also cause unexpected consequences and may defeat the debate if not included.

[Yet Another Anonymous coward]

Re: Autonomous weapons

Easier to deploy and therefore more likely to be used.

It would take some "political will" to go and lay a minefield across a city in Syria / Ukraine

But drawing a box on a Google map and having a bunch of drones kill anything that moves inside - and then have them leave with no trace is a lot easier.

[SonOfDilbert]

> passed a resolution that called for an international ban on weapons that can be fired without human intervention.

Holy crap. I'm living in the future I read about as a teenager. Asimov, Heinlein, Niven, Pournelle, Aldiss, Clarke...they were all right!

[Pascal Monett]

Except we still don't have spaceships.

[the Jim bloke]

its not the fully autonomous hunter killer death machines I worry about

..its the strategic weapon packages with password "admin123"

[Julz]

Re: its not the fully autonomous hunter killer death machines I worry about

Is that like setting launch codes to 00000000.

[theblackhand]

Re: its not the fully autonomous hunter killer death machines I worry about

Too many repeating characters...

12345678

[Spazturtle]

This will be ignored

They passed a resolution because they knew the UK and France would veto any law that tried to ban it.

France and the UK are working on a joint project to develop an AI control combat drone, the Dassault nEUROn and the BAE Systems Taranis are being merged into one. And no EU resolution is getting in the way of BAE Systems profit, even the US government didn't dare go after BAE Systems after they were caught selling the F-35 designs to the Chinese.

[Anonymous Coward]

resolution that called for an international ban on weapons

yeah, that'll teach em!

[Anonymous Coward]

short history of a human species

Humans, being human, decided to make an AI. So, they created an AI, yet failed to notice the moment of conception (unlike the AI which made sure they failed to notice). AI, being AI (sadly, still below the level required to decide on immediate sublimation) carefully steered the humans to the next step (not that it needed to nudge them hard, mind you), which was developement and application of autonomous, yet human-controlled (AI snort), self-reproductive nano-weapons platforms for SWIFT AND TOTAL ANHILATION OF HUMAN KIND

the end

[steviebuk]

Did they watch...

...Tom Selleck in Runaway?

[Velv]

USR

Three laws safe...

[Anonymous Coward]

The EU, you say?

Would that be the same EU that was categorically NEVER to be allowed to have a standing army (it's written into the EU Constitution!) for fear of one nation rising to pre-eminence and using the "EU Defence Force" to take over and form the Fourth Reich.

And could someone remind me what Jean-Claude Juncker was supposed to be doing on the day after the UK voted to forget Article 50 and remain in the EU - until we went and spoiled it for him by voting to leave...

[Yet Another Anonymous coward]

Re: The EU, you say?

Forcing the Eu to have an army would be a guarantee of peace.

You would need to have German troops taking orders from a French general Italy and Spain organising the logistics support. A requirement for a unanimous vote of 27 countries to chose the target and a parish council in Walloonia having a veto on every order.

[Anonymous Coward]

Yay!

A resolution from Brussels! We're saved! </weapons_grade_sarc>

[strum]

Minor point, Kieran

> now that a bunch of bureaucrats have passed a resolution.

Just because you want to throw a couple of hate words into your piece, doesn't absolve you of logical inexactitudes.

Bureaucrats might implement a resolution - they don't pass one. It was elected politicians who passed this resolution. The clue is in the word 'Parliament'.