Inkwell: Authors and Artists

    
We're pleased to welcome P.W. Singer to the Inkwell.


P.W. Singer is a senior fellow at the Brookings Institution and the
director of the institution's 21st Century Defense Initiative. He is
the author of Wired for War: The Robotics Revolution and Conflict in
the Twenty-First Century. For further information, see
http://wiredforwar.pwsinger.com .

Leading the discussion with Peter is Michael Berch. Michael is an
attorney and IT strategist who spent the latter half of the 1980s at
Lawrence Livermore National Laboratory as a consultant on a number of
defense programs, including modernization of DOD's European networks
and redesign of the NORAD air defense network, and was a proponent of
the use and expansion of Internet technology in defense and non-defense
government programs. He remains interested in national security and
defense technology issues.

He was also the founder of the Internet magazine INFOBAHN in 1995,
served as an executive at several Silicon Valley technology companies
including iPIX and Cataphora, and returned to consulting and law
practice with an international focus in 2004.

Welcome, Peter and Michael!
  

    
Thank you for joining us, Peter, and for a thoughtful and absorbing
book.  For those who haven't read it yet, perhaps you could give a
short summary of what it's about, and what led you to this particular
subject? 
  

    
Many thanks! This is my first time jumping into the WELL, so am
excited about it. 

Wired for War is about how something big is going on in war today, and
maybe even in the history of humanity itself. The U.S. Military went
into Iraq with a handful of drones in the air, we now have 7,000. We
went in with zero unmanned ground systems, we now have 12,000. And the
tech term "killer application" doesn't just describe what iPods did to
the music industry. It takes on a whole new meaning when you are
talking about robots armed with everything from machine guns to
missiles. 

But we need to remember that the Packbots and Predator drones of today
are the Model T Fords, the Wright Flyers, compared to what's coming
soon. And, one of the people that I met with was an Air Force three
star general, and he said we'll soon be at "tens of thousands of robots
operating in our conflicts." 

So I thought that was a rather big deal, that something revolutionary
was going on in war and technology, but yet no one was talking about
it, mainly because it seemed too much like science fiction. 

Now, as an aside, I need to be clear here. I'm not talking about a
revolution where you have to worry about the governor of California
showing up at your door a la the Terminator. When historians look at
this period they're going to conclude that we're at a different type of
revolution, a revolution in war, like the invention of the atomic
bomb. But it may be even bigger than that, because our unmanned systems
don't just affect the "how" of war fighting, they affect the "who" of
fighting at it's most fundamental level. That is, every previous
revolution in war be it the machine gun, or be it the atomic bomb, was
about a system that either shot faster, went further, or had a bigger
boom. That's certainly the case with robotics. But they also change not
just the “how” of war, but also the “who” of war, the experience of
the warrior, and even the very identity of the warrior. Another way of
putting this is that humankind's 5,000 year old monopoly on the
fighting of war is breaking down in our very lifetime. 

So, I spent the last several years going around interviewing all the
players in this field, from the robot scientists who build them, to the
science fiction authors who inspired them, to the 19 year old drone
pilots who are fighting from Nevada, to the four star generals who
command them, the civilian politicians who decide when and where they
are used, to the other side of the coin, the Iraqi insurgents to find
out what they think of our robots, and us using them, news editors not
just in the US, but in places like the Middle East and South Asia,
people who have to wrestle with the ethics and laws at places like the
International Red Cross and Human Rights Watch.

And so, akin to a book like Fast Fodo Nation, which I admired greatly,
Wired for War brings together their stories, which I almost always
found amazing, but also sometimes funny, and sometimes scary. But it
also looks at how their experiences point out these broader ripple
effects that are going outwards from this technology onto our wars, our
politics, our society, our law, our ethics, et cetera. So the book is
less so much about the machines than about us. 
  

    
One other aspect that I have iften been asked about, which might be
apt for this group is the very different style of writing for a book
coming out of a pointy-headed scholar who works at a thinktank and at
the time was coordinating defense policy for the Obama campaign. It was
considered a major career risk to research and write a serious book on
what many people insist is science-fiction. I was warned by many that
it would scuttle my career. So I decided that if I was going to risk
it, I may as well double down by writing it in a way I wanted to write
it, to have it read like the books I liked to read. So while it is a
book that tackles serious issues, I wanted it to be received by a
broader audience. The result is a very pop-culture filled book, with
lots of anecdotes, lots of stories and characters, but also hard data
and some deep theory. That is, it blends in everything from battle
stories out of Iraq to philosophers like Thomas Hobbes to chracters
from Star Wars to revealing my own love for the Gilmore Girls.
Hopefully, its a book that doesn't just make you think, but also scares
people at some point, as well as makes people laugh out loud at some
point. 
  

    

You're getting some great reviews. One that caught my eye is one that is
mostly laudatory but does have a couple of minor quibbles:
http://www.amazon.com/review/R1B20UT518228A/ref=cm_cr_rdp_perm So based on
your statement your career risk and the reviews so far, are you feeling
vindicated?

Second question: "Issue-That-Must-Not-Be-Discussed." The reviewer I cited
highlights how senior managment can be clueless at anything below the
"50,000 foot level" where "the rubber meets the clouds". Could you say
something about any feedback you've received from the upper military
echelons from items brought out in your book?
  

    
Thanks for coming to Inkwell, Peter.  I'd like to add my voice to the chorus
of those praising the readability of Wired For War, adding that it's all the
more remarkable that it should be so considering how you at least touched on
practically every conceivable angle.

I grew up on the plains of southern Kansas, embedded in if not exactly part
of the farming culture of the area, which mainly meant the cultivation of
winter wheat.  Two types of machines dominated the landscape, tractors and
(during harvest season) combine harvesters, more usually simply called
combines.  While I never actually got to drive one, I was fascinated by
combines, which perform several related operations simultaneously, mostly
without intervention on the part of the operator.  Even better, many farmers
had examples of previous generations of equipment, collected together in
little rusting museums, so gaining some sense for the evolution of farming
equipment was almost a matter of osmosis.

In high school, because I lacked the hands-on experience of my peers, I
opted for auto mechanics and metals shop in preference to advanced algebra
and trigonometry, else I might have ended up some sort of engineer.  During
this time I also consumed a great deal of science fiction.

In college I eventually settled on a biology major, but was dissatisfied,
and after graduation continued struggling toward something I only later
learned to call a systems perspective, which didn't fall neatly within any
single academic discipline.  At about the same time that I discovered
systems theory I also became deeply fascinated with computing.

During the years that followed, an idea that draws on many of these
experiences gradually took shape in my mind, an idea about the application
of robotics to the detailed management of productive land, currently
represented by <http://cultibotics.blogspot.com/>.

I also host a tiny, until recently moribund corner of The WELL known as
the augmentation & robotics conference <augbot.ind.>.

At a guess, you might find the <attack.> conference interesting.  It began
as a response to 9/11, and has carried on through all of the subsequent
turns of events.

To turn back the clock again, one of the first books I ever read more than
a single time was A.F. Casselberry's "Flying to victory"...

<http://openlibrary.org/b/OL6455853M/Flying-to-victory>

...which could be said to have anticipated the unmanned aircraft of today
by fifty years.

Now for a question, which relates less to the new trend towards robotics
than to the enthusiasm for networking which preceded it.  Are you aware of
any experimentation with giving every squad its own remotely-linked geek,
sharing their voice channel and doing the keyboard entry for them, to act
as a liaison to other aspects of the military, such as the drone pilots,
a filter for information resources, passing along only the most relevant,
a buffer between the troops on the ground and the many distractions that a
networked environment can present, and an advocate when they need backup
immediately?
  

    
Just thought I'd point out that there's a video of Mr. Singer's
excellent talk at the TED conference:

http://www.ted.com/index.php/talks/pw_singer_on_robots_of_war.html
  

    
The response from folks in the military has bene fantastic. I've been
out on book tour (am typing from hotel computer lab in Pittsburgh right
now) and so far have been to every single service academy (just back
from west point on monday), as well as most of the military
professional schools like Naval War College, Army War College, etc.
Just great chance to engage with people who are wrestling with this in
the day to day. 

As some intimated, there might have been an expectation of more
pushback from senior generals for highlighting how disconnected things
had become between the technology troops were using on the ground and
those making decisions. But for the most part I have gotten very good
feedback. Indeed, we've gotten formal plaudits (which are now on the
WiredforWar website) from folks like Gen Schwartz, chief of staff of
Air Force, and Gen Mattis, the Marine now in charge of Joint Forces
command. Also got very nice leters from Gen Petraeus and Adm
Staviridis. My sense is that senior leadership really does want to get
this right, but is overwhelmed by the scope and pace of everything
going on, and so was fairly welcoming to my findings. 

To the other question by John in #5: No, I havent yet heard of
anything like this at the small unit level yet.  You could almost
imagine it as "Geek Squad" for the military. They do have it for the
higher levels. Indeed, there is a story I was told by an officer about
how the battle plans they were prepping for the first gulf war actually
were on a computer that crashed, and so they call up the lead techie
guy in the middle of the night to fly down to pentagon and fix it. He
didnt want to and turned out was able to fix it over the phone . Its
unconfirmed story, but still sorta amusing. Of course, now the techie
helper/buffer you describe would likely be someone outsourced to work
in a call center in south asia, which would create even more amusement.
  

    
My college boyfriend -- whose dad worked for GE in the 1960s designing
Gatling guns -- is now working for iRobot -- the same company that
makes robot vacuums and floor cleaners -- designing robots to find land
mines.
  

    
Interesting book. I haven't gotten all the way through it yet but I'm
greatly enjoying it so far. And, boy, it's a lot of material.

I find it interesting that the fighter jocks (and the Air Force
generally) were disdainful, at least initially, of piloting drones
remotely. As the role of drones increases, are they changing their
tune? It seems to me that before long manned aircraft will become the
exception and not the rule. Do you think this is the way things are
going?
  

    
> Of course, now the techie helper/buffer you describe would likely
> be someone outsourced to work  in a call center in south asia, 
> which would create even more amusement.

Heh!  I really had in mind using soldiers, perhaps people who'd been
wounded too badly to return to regular duty, with the idea that they
would be able to bond with the people on the ground and become an
integral part of the team, although I don't suppose a thick Indian
accent would prevent that from happening.  You'd just need to select
for operators with the ability to individuate and care about their
team, despite only becoming acquainted with them remotely.

I don't think it would play well as a drama/sitcom in the U.S. market,
but it might in the Indian market, where the subtle, frequently
ironic humor among operators wouldn't go to waste.  You could even
drag in religion by having a muslim operator whose reliability was
above question but who would nevertheless, inevitably be surrounded
by trust issues.

On to the next question: I don't have your book with me at the
moment, and don't recall exactly what you said about the military
working with game developers, but I do vaguely recall that in
"Ender's Game" particular arcade games were networked into the
recruiting system somehow.  I'm curious whether some of that is
going on now or planned for the near future, but even more curious
whether the military is making use of gamers to test device designs
that are still on the drawing board.  Can you say anything about
that?
  

    
Just a reminder - anyone reading this via the Well's web site who is
not a Well member can pose a question for Peter by emailing me at
mcb@well.com and I'll post it here. 
  

    
Peter, another, more basic question: why robots?
  

    
Sorry to load you up with questions, Peter, but here's another topic that
I've been waiting to ask about...

On page 87 you mention "triggered decay" of nuclear isomers* as a
potential power source for "long-duration unmanned aircraft", referencing
"Weapons Grade" by David Hambling (p152), which also describes bombs
far more powerful than chemical explosives, based on the same effect.

I initially misunderstood this, conflating "isomer" and "isotope", but
have that part straight now.

Given the destructive potential of weapons based on triggered nuclear
isomers, and the fact that they would have an effect similar to neutron
bombs, since the stored energy is released in the form of gamma radiation,
killing without creating a great deal of physical destruction, could you
get behind a move to ban such weapons?

On the other hand, can you imagine any commercial application for
triggered nuclear isomers that would be safe - in terms of both protection
from the radiation produced and security to prevent the material from
being diverted into weapons - and economically viable, given that creating
the material in the first place is likely to be energetically inefficient?

Powering unmanned, long-duration, high-value assets seems like the perfect
match, since neither the radiation nor that fact that it might take 100KWh
in to produce 1KWh out would be of much concern.

* For the benefit of others reading this, note that "nuclear isomer" is
not equivalent to "nuclear isotope", although particular isotopes would
undoubtedly be involved in any instance of creating and triggering the
decay of a nuclear isomer.  Note also that this is essentially battery
technology, not an energy source.  If triggering could be used to
accelerate the natural decay of one isotope into another, that might be
an energy source.
  

    
Is this the purported, bogus "hafnium bomb"?
  

    
Hafnium 178 is one of the isotopes into which energy can be stored by the
creation of an isomer, and from which that energy can be released, either
in a controlled manner or all at once, yes.

See  http://tinyurl.com/coqfjn  (Google Book Search link of Hambling ref)
  

    
My gosh, just back from a friend's wedding and tons to reply to. 

Why Robots (#12)? Well because robots are, as I write in the opening
to the book, Frakin cool (an homage to Battlestar Gallactica). the
longer answer goes back to my initial posting, that we are living
through this remarkable revolution, where these amazing technologies
are starting to have great impact on our wars, our politics, etc. and
yet no one was writing or talking about it. So, besides the incredible
stories I could collect, this is what motivated me, to capture a
historic moment in time, and maybe even offer a framework of
understanding for moving forward. 

Ender's Game and recruiting via Video Games (#10): Yes, indeed, this
is another one of those areas where sci fi is starting to play forward.
I actually had a draft chapter entitled "Meet the Sims --And Kill
Them," which looked at the growing connections between the video game
industry and the military. This is not just the manner in which the
military uses more and more video game technology, free riding off of
the research work by companies like Sony and MSFT, as well as how they
have already trained up a generation in their use. But moreso, it
looked at the amazing story of "America's Army," which was a video game
developed for the XBox to help with recruiting and has become a huge
best seller. Alas, the book was already a bit long, so we had to chop
that section out. 

Are USAF attitudes changing on drones (#9)? Somewhat but slowly. The
senior leadership, which had replaced a previous set who were fired in
part for not meeting Gates' push for more Predator flights, are doing
better and ramping up buy rates and pilot training for these systems.
But it still is somewhat begrudgingly and not well integrated with rest
of the force. And indeed, there is talk of creating a new air wing
that would focus just on counterinsurgency missions. But, of course,
the only plane the leadership talked about buying for it was a light,
manned, propellor plane (a back to the future mentality), never once
exploring why it would be better than the unmanned systems. Within the
force, its mixed as well. There is more and more experience being built
up with these systems, so greater constituency, but you still see
resistance. I was recently at US Air Force Academy and one of the young
officers there was soon to join a Predator squadron. And he was sorely
disappointed about it, describing how he wanted to fly and instead his
new job was "going to be boring." So, while his unit will likely
engage in more combat and critical missions to American national
seucirty than almost any other in the force, it is still "boring" to
him and he wasthe object of his mates jokes. I told him there was
likely some young officer in 1919 at West Point who equally was
probably disappointed to be joining a tank unit instead of the more
exciting and prestigious horse cavalry units. 

Isomer question (13): I am going to dodge this one. Hambling really is
a far better expert than I on it, so anything I give would be more
speculation than fact based. I would say that as weapons move from
theoretic to real, we have an important ethical responsibility to weigh
their utility against the potential harms, intentional or
unintentional. If we just push forward, without any consideration, we
are simply repeating the mistakes of the scientists back in the 1940s,
who built the atomic bomb and then afterwards regretted it. 
  

    
That warning would seem to apply even in the seemingly innocuous case
of my own vision for a landscape in which tractors and tillage have been
replaced with more intelligent, intricate machines performing gardening
techniques.

One page 247, you quote Qiao and Wang, authors of "Unrestricted Warfare"
as saying "We believe that some morning people will awake to discover
with surprise that quite a few gentle and kind things have begun to have
offensive and lethal characteristics."

That made me think.  What would military planners and the technological
wizards of DARPA and other such institutions see in the prospect of a
landscape full of intelligent, networked machines?  Ubiquitous
surveillance, certainly, but would it stop there?  Would they insist
upon back-door access to systems sold into unfriendly markets?

While a Skynet-gone-berserk scenario may be too improbable to bother
seriously considering, malicious repurposing of commonplace technology
by remote crackers is all too likely, including as part of coordinated
attacks.  Whatever is connected to the internet and somehow effects the
physical world might be made an instrument for mischief, or worse.
  

    
without the arrow and the bullet, fighting was a face to face affair

shelling and bombing enabled killing at a distance

"As Anthony Lewis once wrote, our military technology is so advanced
that we kill at a distance and insulate our consciences by the
remoteness of the killing."
http://en.wikipedia.org/wiki/Gloria_Emerson
  

    
For those viewing this topic from inside The WELL, I've started a topic
in the Science conference for continuation of the Hafnium debate, and 
related subjects.

<science.688> EM manipulation of atomic nuclei: Hafnium...
  

    
Perhaps the best expression of the pastoral cybernetic aesthetic
expressed in Richard Brautigan's poem "All watched over by machines
of loving grace" is that found in Asimov's "The Naked Sun" on the
planet Solaria.  But taking the phrase "machines of loving grace"
by itself, the robots found in the frozen conclusion of Spielberg's
film "A.I. Artificial Intelligence" are the best.

If we could build machines like those, maybe they could serve as
peacekeepers.
  

    
Peter,

The most interesting chapters for me were on autonomy ("Always in the
Loop?") and international law and human rights ("Digitzing the Laws of
War"), which to me at least are closely related.  While right now it
appears that goals and rules are communicated to robotic devices at a
pretty low level, or are teleoperated by humans, that clearly won't
always be the case.

Since it doesn't look like Asimov's Three Laws of Robotics -- which
you're careful to point out are fictional -- are likely to be the the
basis of high-level rules for robots, what might be instead? And who
would decide?  I can imagine a military officer thrilled with the idea
of telling some unmanned weapons platforms, "take that ridge", but what
he or she really means is, "take that ridge, but only engage obvious
combatants, and destroy only the structures you have to, and make sure
not to harm the nearby school or mosque."  It's hard enough to do that
with human soldiers, but it seem to be an order of magnitude harder
with robots, even if they are able to comprehend the high-level
orders.
  

    
Michael, 

Yes, exactly, that is the challenge. Here is a thing I did for IO-9 on
the issue of Asimov's Laws:
http://io9.com/5187218/the-3-laws-may-not-be-enough-to-guide-robot-warriors

In the book, I go into this is FAR more depth, but I think that we are
going to have to turn back to the original pillars of the laws of war
that are supposed to guide whether a system can be judged as legal or
not, such as their ability to discriminate between lawful and unlawful
targets,  etc. We have been ignoring them with our latest generation of
technology. But there is a last one of these pillars I sense we may
soon hear more of, whether society finds them objectionable in some
way. This is why blinding lasers for example are banned, and was also
the driver behind ban of bioweapons, even though they might be
militarily useful. But again, this is why we need to engage with the
technology, rather wait until after the fact, akin to what happened
with land mines. As well, it is critical under the law that we ensure
accountability, that there is a clear chain of responsibility for who
to turn to when things go awry. This is something that many roboticists
don't want to talk about (and one recently got angry with me for even
suggesting it in the book), but I think is imperative here. If we can't
establish that clear accountability that I believe extends from the
user in the field all the way back to the inventor, so that we can find
out just exactly where things went wrong, rather than putting all the
blame on the end user, then again, the system may not be one we ought
to use. 
  

    
Just to bring in another piece of fiction - set for release May 21 is
Terminator Salvation, the prequel, in which Skynet becomes self-aware
and destroys most of humanity, starting a war between machine
intelligence and the survivors. Are you working on an op-ed for the
occasion? Perhaps you could start sharing some thoughts about the
extremes here.
  

    

It strikes me from that "3 laws" article that there significant boundry is
where "robots" start being autonomous. Given my work background in
computer software, I get really, really, REALLY nervous about bugs and
design flaws. It's bad enough when we have "ordinary" bugs but at least a
human can take action. But an autonomous robot with a bug is quite
something else.

Typically designers and developers are overly optimistic about the odds of
their work being perfect, but I'd want a "paranoid" in charge of such
projects because a paranoid would suspect that things are wrong and they
most likely will be.

The Roomba is a good example: 97,500 hits on "Roomba bug" and 55,300
on hacking Roomba. That's enough evidence for me.
  

    
One thing that struck me with all this automation is that if going
into battle starts to become only marginally more dangerous than, say,
driving on the freeway, wouldn't that tend to make it easier to resort
to force. Why bother with the complexities of non-violent solutions
when force can be used so quickly, effectively and safely?