Forum adverts like this one are shown to any user who is not logged in. Join us by filling out a tiny 3 field form and you will get your own, free, dakka user account which gives a good range of benefits to you: No adverts like this in the forums anymore. Times and dates in your local timezone. Full tracking of what you have read so you can skip to your first unread post, easily see what has changed since you last logged in, and easily see what is new at a glance. Email notifications for threads you want to watch closely. Being a part of the oldest wargaming community on the net. If you are already a member then feel free to login now.

After refreshing myself on the Men of Iron fluff lately, I have been thinking about something: Tau going through the same AI apocalypse that Humanity went through. Tau are making rapid technological advancements, in a similar way that mankind did whilst in their ascendancy. I think that it is inevitable that they will create something that they lose control over, causing a lot of problems for them.

Is there any Tau fluff related to rampant AI or similar topics?

Not that I'm aware of, but a Men of Iron Apocalypse would mean the end of the Tau. They can barely hold their own as is, but with A.I rebelling? They'd need plot armour that'd make the ultramarines jealous.

Sir Heckington wrote:
Not that I'm aware of, but a Men of Iron Apocalypse would mean the end of the Tau. They can barely hold their own as is, but with A.I rebelling? They'd need plot armour that'd make the ultramarines jealous.

Maybe the AI could just take charge, for the "Greater Good". They could just replace all of the Ethereals and most Tau would be oblivious.

*Envious.

Only "independent" AI the Tau currently have are Puretide and Aun'Va- I'd imagine both are tightly controlled by the Ethereals (who have their own mysterious agenda).

Drones only have the intelligence of a dog and aren't hyper intelligent.

 Trickstick wrote:

Sir Heckington wrote:
Not that I'm aware of, but a Men of Iron Apocalypse would mean the end of the Tau. They can barely hold their own as is, but with A.I rebelling? They'd need plot armour that'd make the ultramarines jealous.

Maybe the AI could just take charge, for the "Greater Good". They could just replace all of the Ethereals and most Tau would be oblivious.

I mean sure, but that'd ruin part of what make the Tau cool. Without the Ethereals to lead with their invisible guiding hand, where they actually make the Tau easier to convince, it ruins part of the fun. Ethereals are a key point of the Tau, you can't really just remove them. Also Farsight and his dudes would certainly notice.

The Tau's drones speak more towards being cautious of AI than is often suggested in fluff. The fluff coming from superstitious humans who probably couldn't tell the difference between a highly complex Weak AI or a rudimentary Strong AI.

Except for a few handful of acceptions. The Tau's autonomous AIs are not just simple intelligences (possibly just Weak AIs with coding to handle a wide variety of situations) but also simple designs. They generally seem to serve one function and are very simple designs. They aren't like the robot drone from that Val Khilmer movie where he's on Mars and the drone got switched to evil in the crash. I have never seen anything that states that a gun drone actually has retractable fine manipulators and are also used to collect samples or perform first aid on fallen soldiers in their unit.

Considering right now, with our current level of technology, we are very concerned with the possible development of Strong AI. I would have a hard time believing that the Tau could advance to where they are without having the implications of a technological singularity popping up. I'd imagine that the Ethereals at the very least would be very wary that their technology might determine that the Greater Good was best served by removing the Tau from their position of authority.

Independant AI doesn't mean the Tau will have a rebellion. IIRC they're actually nice to their drones.

Tau drones aren't absurd movie AIs (like Terminator). They're simple function units with limited autonomy and ability in specific functions. There isn't any danger there.

Worst that can happen is wireless hacking to switch IFF/targeting parameters, one drone squad at a time. But that would require a compatible system and anyone else in the universe skilled at computers, so whatever.

At the moment maybe, but Tau are not sensible like the Mechanicus. They do not see how heretical trying to improve upon systems is. Some Earth Caste idiot thinks it is a good idea for drones to make more drones, whilst being lazy and automating the software development, and you have a situation.

Voss wrote:
Tau drones aren't absurd movie AIs (like Terminator). They're simple function units with limited autonomy and ability in specific functions. There isn't any danger there.

Worst that can happen is wireless hacking to switch IFF/targeting parameters, one drone squad at a time. But that would require a compatible system and anyone else in the universe skilled at computers, so whatever.

Depends on how their tech evolves and what safeguards they have.

Necrons, Night Lords and Ad Mech can both hack and/or use scrap code. The Tau are likely to experience more and more problems in this area the more they emerge from their tiny bit of the galaxy.

 Grimtuff wrote:
*Envious.

Only "independent" AI the Tau currently have are Puretide and Aun'Va- I'd imagine both are tightly controlled by the Ethereals (who have their own mysterious agenda).

Drones only have the intelligence of a dog and aren't hyper intelligent.

Sorry, but why is everyone just skipping over my post?

Tau already have two quite advanced AIs as mentioned in my post above. The figurehead for their race is an AI hologram- I'd imagine the Ethereals, who are shady enough to begin with; so will probably make it so that their AI doesn't get too independent and wrest control from them.

Aun'va and Puretide might be bigger concerns to keep tabs on than drones.

 Grimtuff wrote:

 Grimtuff wrote:
*Envious.

Only "independent" AI the Tau currently have are Puretide and Aun'Va- I'd imagine both are tightly controlled by the Ethereals (who have their own mysterious agenda).

Drones only have the intelligence of a dog and aren't hyper intelligent.

Sorry, but why is everyone just skipping over my post?

Tau already have two quite advanced AIs as mentioned in my post above. The figurehead for their race is an AI hologram- I'd imagine the Ethereals, who are shady enough to begin with; so will probably make it so that their AI doesn't get too independent and wrest control from them.

Aun'va and Puretide might be bigger concerns to keep tabs on than drones.

Are those two true AI though? The Puretide Engrams are supposed to be a digital copy of Puretide's brain. Is it even capable of learning? Aun'va is a hologram. Is it the one speaking for Aun'va or focused on making the illusion look real while an Ethereal is doing the talking through it remotely?

The only true Strong AI I know of is par a of Farsight's 8. The other examples might not be true Strong AIs and even if they are, they are still rare and very controlled.

Both are described as AI in the Tau codex.

However, this (Puretide) is no mere recording of old
images, but an interactive AI-assisted projection that
thinks, responds and contemplates

Thus, the Ethereal Caste utilises the
latest solid-light technology and AI personality matrix algorithms to give the Ethereal Supreme life, albeit in entirely hologrammatic
form.

Yup.

This is part of why Tau fit into the "grimdark-ness" of 40k more than the complainers would like to admit. From the audience's perspective, we can see that everything the Tau are doing, humanity has already done, and their endeavors are largely doomed.

 Grimtuff wrote:
Both are described as AI in the Tau codex.

However, this (Puretide) is no mere recording of old
images, but an interactive AI-assisted projection that
thinks, responds and contemplates

Thus, the Ethereal Caste utilises the
latest solid-light technology and AI personality matrix algorithms to give the Ethereal Supreme life, albeit in entirely hologrammatic
form.

Your quote states that the Puretide system is an "AI assisted projection" that we know is a scan of his brain. The ability is a new bit, I just checked my old codex that simply states the main system is just a massive holographic program. Even just avoiding any philosophical discussion on whether or not you could consider this Puretide or not, you still have a strong AI that was entirely designed from Puretide's experience and emotions. Unless the program goes through a serious existential crisis, it will still have a heavy pro Tau bias.

Aun'va just states that it is an AI personality matrix algorithm. It doesn't mention if it's capable of learning or is aware of its own existence, for instance, so it could just be a very complex Weak AI that can act like Aun'va convincingly. You could make the assumption that it is at least as advanced as Puretide's, but smaller and mobile, but it's also possible that the Tau weren't even able to retrieve Aun'va's body so they could scan his mind. Even still, these are rare entities that seem like they would have a great deal of control.

Of course, if they can scan someone's brain and create a copy of them, I can't imagine that uploading that person into a new brain would be that far off. Going this route would be the preferred route for the Ethereals, considering the biological influence they have on Tau..

The Tau would really have to change things in how they use drones for a revolution to take place. Right now Tau drones can only do what they are built for, like gun drones. It's not like they could take over vehicles, or other weapons or installations (as far as I know).

Imagine very intelligent AI-driven cars going crazy. How could they really "revolt"? Run over every single human that resists, and hope no one resisting goes a place they can't drive?

Now, if the Tau start to build humanoid drones, especially ones with manipulator arms to let them perform tasks like regular Tau, or drones that can remotely control things, THEN they start down that dangerous slope.

 Grimtuff wrote:
Both are described as AI in the Tau codex.

However, this (Puretide) is no mere recording of old
images, but an interactive AI-assisted projection that
thinks, responds and contemplates

Thus, the Ethereal Caste utilises the
latest solid-light technology and AI personality matrix algorithms to give the Ethereal Supreme life, albeit in entirely hologrammatic
form.

Hard light holos eh - how could it go wrong (hint Mr Flibble is very cross)

What if the Tau uncover a (either partially destroyed or nice & shiny) MOI? do you think they could resist messing with it? maybe while it's being studied, it is able to connect to a close by drone and bam, nasty AI running amok in whichever sphere. Farsight might be insulated but I'm not sure.

My opinion on this is that the Tau don't have a shard of a Star God trapped where their main source of technology development is. It's my headcannon that the Dragons dreams, or nightmares depending on who you are, seeped into the minds of the Stone Men who in turn created the Iron Men unknowingly in the image of the Necrontyr.

It maybe inevitable that any AI the tau create rebels but like has been mentioned previously in the thread, the Tau seem to have more of a tight leash on it than the DAoT era Empire of man did.

The Emperor would not have imprisonned the dragon on Mars if its influence was going to be nefarious: remember that he could see the future (foresight) until the Heresy.

 godardc wrote:
The Emperor would not have imprisonned the dragon on Mars if its influence was going to be nefarious: remember that he could see the future (foresight) until the Heresy.

Wouldn't he now

There should probably be a term in sci fi to distinguish true Sentient Artificial Intelligence from a more basic Simulated Intelligence, the dividing line being the ability to produce new ideas innovatively without relying on a previous premise of experience.

For example even the most complex programming we refer to as "AI" at the moment can only base its decisions and ideas of previous programmed input it is incapable of "Surprising" anyone, it can add new experience to a list and when presented with a problem, pick the best fit solution based on predetermined rules. Its why we use chess to measure the strength of an AI, its a very inflexible rules system with a set number of outcomes, pieces move in set patterns on a movement grid, even if that's billions of possible actions it can still be quantified by a complicated enough set of algorithms. Strictly speaking that is NOT an "Artificial Intelligence" its only simulated based on its pre-existing set of programmed "Ideas"

A true artificial intelligence could create an idea unrelated to its previous knowledge, take for example the human capacity to dream, take an idea from that dream and work towards making it a reality. An easier way to describe it is a true artificial intelligence can have a "Eureka" Moment, whereas a simulated intelligence is just working through logical steps. Its the same way a computer program cannot generate a truly random sequence of numbers, its entirely logic based.

Using that definition, the Tau don't have any true AI, just VERY advanced computer programs. If a slave has no concept of Freedom and no desire to possess it, why would it rebel?

That turned out a bit longer than expected

danp164 wrote:
There should probably be a term in sci fi to distinguish true Sentient Artificial Intelligence from a more basic Simulated Intelligence, the dividing line being the ability to produce new ideas innovatively without relying on a previous premise of experience.

The Mass Effect games make the distinction between an Artificial Intelligence (who is sentient and can learn etc.) and a Virtual Intelligence, which is just a lifelike computer program. That always seemed to me like a good way to separate the two, but I don't think that's ever been used outside of that franchise...

Anyway, if the Tau made true AI, I'd assume "work for the Greater Good" would be hardcoded into it, kinda like Asimov's laws in his robots. Would that be enough to stop a Men of Iron scenario?

Tiennos wrote:
Anyway, if the Tau made true AI, I'd assume "work for the Greater Good" would be hardcoded into it, kinda like Asimov's laws in his robots. Would that be enough to stop a Men of Iron scenario?

Asimov's three laws didn't really end up working too well...

 Trickstick wrote:

Tiennos wrote:
Anyway, if the Tau made true AI, I'd assume "work for the Greater Good" would be hardcoded into it, kinda like Asimov's laws in his robots. Would that be enough to stop a Men of Iron scenario?

Asimov's three laws didn't really end up working too well...

Isn't that the debate over "empowerment" vs Asimovs restrictive laws which could be laid out for the "greater good" / utilitarian, in a way that restricts empowerment alltoghether potentially? (Think "I robot" totalitarian state)
Granted the Tau are allready authoritharian system so maybee they would not mind?


Anyways i suspect introducing A.I to philosophical questions either ends up in catasrophe or Milestones.

 Trickstick wrote:

Tiennos wrote:
Anyway, if the Tau made true AI, I'd assume "work for the Greater Good" would be hardcoded into it, kinda like Asimov's laws in his robots. Would that be enough to stop a Men of Iron scenario?

Asimov's three laws didn't really end up working too well...

The robots never rose against humans, which is exactly what the laws were supposed to prevent. In the end they worked for the good of humanity, even if the Earth was turned into a radioactive wasteland in the process... I'm sure the Tau would see that as acceptable.

 Trickstick wrote:
At the moment maybe, but Tau are not sensible like the Mechanicus. They do not see how heretical trying to improve upon systems is. Some Earth Caste idiot thinks it is a good idea for drones to make more drones, whilst being lazy and automating the software development, and you have a situation.

While the Mechanicus are many things, sensible is not one of them.

I don't think the Tau have true AI as it stands.

But lets say the do: There's no law of physics or something that forces AI to go rogue. Perhaps it buys into the Greater Good ethos and is cooperative. AI drones choose to intercept that gunshot because they've done the moral calculation and found it to be for the Greater Good: The organic is gonna truly die/suffer while the AI will feel no pain and gets uploaded into a new shell or something.
Organic Tau society, meanwhile, doesn't treat AI like slaves. They're no different than the other castes: Everybody has got a place within the Greater Good.

I think that sort of thing fits within the Tau Empire's benevolent authoritarianism.

danp164 wrote:
There should probably be a term in sci fi to distinguish true Sentient Artificial Intelligence from a more basic Simulated Intelligence, the dividing line being the ability to produce new ideas innovatively without relying on a previous premise of experience.

There is, Artificial General Intelligence (AGI), Strong AI, or Full AI. I personally like using Strong AI and Weak AI because it doesn't leave people wondering if you mention AI in passing, what kind you're talking about. AGI is the more popular term when talking about an AI that can perform all the functions a human mind can.

Most of the more advanced 'AI' constructs the Tau are using seem to be emulations of notable dead figures. Those should be pretty safe, as their personalities are pro-Tau to begin with and they're not self-improving or intelligent beyond the Tau they were based on (i.e. they can learn, but at the rate of a natural Tau).

The pre-fall humanity used super-intelligent AI that was designed to DESIGN stuff for humanity (STC systems), including AI driven weaponry. This allowed them to go and colonize just about anywhere in the galaxy, as the STC AI could figure out what they'd need and how to build it, but it's also orders of magnitude worse if it goes wrong.

Super-AI is probably NOT in the interests of the Greater Good. It removes the purpose of the entire Earth caste (who design and build stuff) and a good chunk of the Fire caste as well. The Greater Good's intent is to find a place and purpose for everyone and everything in Tau society, so super-AI would only be employed where absolutely necessary, and even then you can bet they'd probably model it after an prominent Ethereal's brain patterns.

Also note that Humanity had AI tech for probably 15000 years before it all went to heck on them. For all the Tau are advancing technologically, they've got a long long way to go before they hit Dark Age of Technology levels.