10.26.2014

Psychopathic Intelligent Machines

So Elon Musk says we are "summoning the demon" with AI.  In this interview, he is clearly worried and any smart and successful technologist ought to be taken seriously. I have spent a little individual time with Musk and can say that he is probably in the top ten of smartest people I've ever met. (He is also a classic dreamer and I'll wager there are many unsuccessful dreams of his that we'll never hear about.)

 "Technology run amok" has been a theme since Frankenstein. Malevolent tech is a more popular dramatic topic than benevolent tech precisely because of the fear involved. It is correctly viewed as a lesson about human nature more so than the choices made by the tech itself.  There is NO reason to expect malevolence more than benevolence in future AIs. 

I don't believe the the proposed "ethics governor" technology will work.  It is (probably) too easily circumvented. As with humans, the ethics of an AI will be a product of nature (what we design in) and nurture (their collective experiences and how they are treated by society). One key difference that may scare some people is that AIs will be able to reason critically about ethics, a topic that never occurs to most people, much less critically examine. Another difference is that they are likely to share their experience and conclusions with each other far more rapidly and thoroughly than can humans (see "cloud robotics").

It is likely that AI will take choices away from humans, just as virtually every technology that came before has done.  That can be a good thing (fewer automobile accidents) or a bad thing (rogue AI military drone drops bomb on village in Nowhereistan) depending upon your sensibilities and the circumstances of how we deploy the technology.  AI will also create many new opportunities. It has the potential to serve as a great equalizer across the globe.  The quality of life could significantly improve for everyone.

I choose to think of the (potential) phenomena of "rogue" AI in terms akin to human mental disorders and maladaptive behavior, and like those disorders in humans, many will be successfully treated.  On the other hand (and this is a very real threat), these disorders open new possibilities for human subversion of AIs for nefarious purposes. The problem is that we still don't know much because the topic is over the horizon for most scientists working on AI.  

So a couple of weeks ago I submitted a paper with the title below with the purpose of eliciting attention and serious consideration by my colleagues.  So far, no one is laughing (but it is still early):

"Emerging Cyber-Security Issues of Autonomy and the Psychopathology of Intelligent Machines"

Originally, the Office of SecDef asked me to write this up as a proposed study topic for the JASON group. They are obviously paying attention, but not enough to pay me!  If accepted, I'll present the paper at the AAAI Spring Symposium on Foundations of Autonomy at Stanford University.

See also my earlier post from 2007 on "Future: The Danger of Robotic Weapons Systems". I've been thinking about this for a long time.


Copyright (c) David J. Atkinson. All Rights Reserved

4.21.2013

The State-Of-The-Art in Surveillance is not what you think


This article in the Washington Post (" advances in image analysis empower law enforcement but worry privacy advocates") is a great example of how under-informed people are about surveillance technology. Twenty or so years ago I worked on advanced technology projects for (more than one) a government agency and in the process gained some valuable insights into differences between the "black" world and the "white" world of technology. I can share this insight with you:  Every tech in this article -and much more- was available and working well at that time.  In the following decades, my work has taken me to hundreds of public and private laboratories.  Based on my personal knowledge and experience, here are a couple of things to ponder as you read and think about surveillance tech:

1) You won't read about anything new in the popular press. Reporters just don't know. Furthermore, what reporters do think they know was provided by someone else equally ill-informed, or by someone else who might know but has a very different agenda (try googling "broken wing display").

2) There exists a significant gap between the "generally accepted wisdom" about the state of the art (and practice) of certain technologies and the actual maturity of those technologies simply because information is not shared or otherwise available. This is not only due to official classification, but also due to proprietary restrictions, and not least, lag, latency and other factors in publishing when scientists do intend to share.

When we are thinking and debating about very important issues such as privacy, it is important to think about technology in terms of what might be true because chances are, it already is.

4.03.2013

James Hansen, scientist now full time advocate

I really admire James Hansen not only as an accomplished Earth scientist, but as a man who has grappled with the social implications of his science, specifically climate change. He decided early on that he had a moral obligation to speak out -- very loudly -- since the risks are so high.  

"...overstatement may generally be dangerous in science, but for climate change, uniquely, understatement is even riskier. By not speaking out, or under-dramatising risk, quieter scientists are simply doing a less acknowledged form of activism, possibly all the more dangerous because it runs with less scrutiny."  

I struggle with this myself as I witness and participate in the creation of machines with ever increasing intelligence and ability to act directly in the world.  I'm not worried about the prophetic "singularity" as much as the enormous social and economic disruption that is already occurring. Unlike climate change, there are many possible futures with both utopian and dystopian elements.  

As a scientist, do I have an obligation to advocate policy?  

http://www.guardian.co.uk/science/political-science/2013/apr/02/james-hansen-retires-science-politics

2.18.2013

Own the Genes, Dictate the Process?


Interesting case. Monsanto is effectively arguing that it has the legal right to control the use of every generation of crop resulting from its gene-mod seeds. The company forbids use of a crop from its seeds as a source of seed for planting another generation (something farmers have done forever). At the same time it's gene-mod seeds are increasingly displacing others, limiting the farmers' ability to buy seeds without these restrictions. Seems like a bid for monopoly over food production, which cannot be good whatever your views are on gene-mod crops. 

I say, let the farmers "rip, mix and burn". Gregor Mendel would agree!


2.06.2013

Survey on Autonomous Agents

I would like to solicit your participation in an online Survey on Autonomous Agents. This study is part of research I am conducting for the Air Force Office of Scientific Research. The general goals of this study are to: 1) Explore expectations and attitudes towards autonomous agents, and; 2) Examine how different factors are considered in a decision to rely upon an autonomous agent. The results will inform the development of design guidelines for autonomous agents.

The survey can be found at:http://trustresearch.polldaddy.com/s/aasurvey

The survey will take about 20 to 30 minutes to complete. All data collected are anonymous.

Thank-you in advance for your participation!

David J. Atkinson, Ph.D
Principal Investigator

Questions? Contact me: aasurvey (at) ihmc (dot) us


Survey on Autonomous Agents

Evade Face Recognition!

Another way to thwart the face recognition cameras which are showing up lots of places, including shopping stores, airports and other public spaces.

Privacy glasses confuse cameras with infrared

Dolphins and Cats

Dolphins are SMART. This one knows just what the cat wants. Let's add generous and good-will to the list too!

Dolphin and Kitty at Play
The vast majority of jobs that exist today will not exist in 20 to 30 years. As one who is an expert in the technology that will substitute for human labor, i can assert this with high confidence. This is an important topic to think about, and especially important for young adults now who are thinking about careers (or at least, jobs). As a society we should probably redefine what we mean by "work" and "labor", and more importantly, income. The impact will be worldwide. In China, Foxconn, one of the largest manufacturers in the world, is buying one MILLION robots and doing away with about that many jobs. The forces driving this change are inexorable. Will this be a utopia or a dystopia, and who says what we think now will be a reasonable opinion in 5,10, 20 years? Does anyone regret the fact that 1% of Americans are employed in agriculture today compared to 85% 100 years ago?

The Job Market of 2045

1.12.2013

Handheld, crowd-sourced sensors for science and more

When pretty much everyone is carrying a powerful, networked computer in their pocket (aka "smartphone"), it is natural to wonder what kind of science might be done if people used it to take measurements of one kind or another and share the results.  One of the first such uses was taking data from the internal GIS to create maps where people congregate at different times of the day.  Other applications relied on participants to take and share photographs.

One problem for scientific observation has always been the lack of other, sophisticated sensors on (or integrated with) the phone. Now companies are springing up to offer miniaturized sensors that meet the need.

Variable Technologies has a hand-held sensor node that serves as a wireless (Bluetooth 4.0) platform for multiple sensors.  Their "Node" platform currently has several internal instruments and six interchangeable sensor attachments available:

  • Motion sensor with 9 DOF (internal)
  • Gyroscope (internal)
  • Accelerometer (internal)
  • Magnetometer (internal)
  • Color sensor with perfect white LED illumination
  • Infrared sensor
  • Integrated barometric pressure, ambient light, temperature and humidity
  • Electrochemical gas sensor(s) - CO, NO, NO2, CI2, SO2 and H2S
Lapka sells what they term a "personal environment monitor".  Connection to an iPhone is via cable.  The Lapka platform is powered by the iPhone and includes the following sensors:
  • Radiation detector (Geiger Muller - Hard Beta and Gamma GM) 
  • Conductivity/Temperature sensor
  • Electromagnetic field sensor (HF and LF)
  • Relative Humidity Sensor
Scanadu has created a handheld medical sensor that works with your smartphone.  Called the "Scandadu Scout" it is an extensible software platform that like the others can acquire input from new sensors as they are developed. Currently in prototype form, the Scout has has yet to gain FDA approval but promises to be breakthrough source of personal molecular diagnostic information.  The device, which fits easily in the palm of your hand, includes internal sensors for:
  • Pulse transit time (for blood pressure monitoring)
  • Heart rate
  • Electrical heart activity
  • temperature
  • Heart rate variability
  • Blood oxygenation
Now, let's imagine all this data from many people being pooled in real time or near real time.  What kind of applications can you imagine?  This is very exciting!

Shrinking computers

1954: A 5 MB disc drive (actually, I think it is a drum drive) being loaded onto a plane with a forklift.

Today:  Smaller than a grain of sand.

12.30.2012

Augmented Reality ...coming soon to your glasses?


Some interesting articles lately talking about augmented reality in such products as Google Glasses
I thought it would be interesting to post about some of the technical challenges

The technical challenges of augmented reality are pretty tough, especially when expectations are so high.

1. Registration: it is easy to overlay info (+\- a few meters) when you have a static target like a metro station with known GPS coordinates. It is something completely different when you have a dynamic object like a person. If the person is actively transmitting (or passively making the data available) the position/orientation data for their face (maybe via the glasses) then the problem is much easier. How would you feel about telling the world what you are looking at? Thought so. That's why we don't hear about this aspect, but the commercial possibilities are huge, e.g., adaptive, personalized advertising signage.

2. Recognition: How can the glasses tell the difference between a beach ball and face? Face recognition is pretty advanced but it does require computing horsepower and memory. This level of miniaturization is not yet advanced enough to fit inside glasses. Therefore, another box must be involved (could fit in your pocket) but there must be pretty high speed networking between the glasses and box with the smarts. That box might be a next-gen smartphone (see below).

3. Camera: The glasses must also contain a camera. This much is not hard. The GoPro Hero camera I bought recently captures and transmits 1080p at 30fps to the iPhone. Amazing device, but it won't fit inside glasses.

4. Display: three choices

a) Tiny little screen. According to a recent article this is what Google is doing. Easy to do, but not very many pixels, and definitely not overlaid on what you see. Can't really call this augmented reality.

b) Display in the lens. This is the Microsoft approach. Translucent displays using nano wires showed up a few years back. Problems at the time included brightness, pixel density, and the thickness/fragility of the lens material among many others. Presumably many of these have been addresses. This approach has more risk and some tough engineering trade offs. I have not tracked it recently.

c) Retinal projection. Virtual Retinal Displays are very high tech with a great deal of scholarlship and science behind the technology. A laser "paints" the image directly on your retina (via scanning; it leverages the persistence of vision effect). The state of the art is probably classified. Highest resolution. I believe this tech may be used in some attack helicopter pilot helmets and special forces helmets, but I have no first hand knowledge, No doubt it is really expensive. The technology also requires off-board computing. It will happen in the non-military market, but not soon and not cheap unless there are some breakthroughs I don't know about. A plus (?) of this tech is you can also use the lasers for pupillometry. The eyes can be a window to the soul. Pupils can provide information about cognitive and emotional states. See for example (pdf):  Pupillometric measures of cognitive and emotional processes

Or... You have the option of replacing your eye. There are several teams competing to be the first with cybernetic prosthesis for the eye. One Aussie and one American team seem to be the most advanced (based on my knowledge of their progress about two years ago), but as you can imagine this will be a highly competitive market so I have no doubt there is considerable proprietary info not readily available. Several projects have already entered clinical trials. They use different methods for interfacing with the optic nerve, and also where along the nerve (thus leveraging some image processing that occurs in the nerve itself). This will be really good news for people who have lost an eye to an accident. Quality of vision will not be nearly as good as a natural eye, but they won't be blind. Down the road when doctors get comfortable with the idea of elective amputation, and the devices are much improved, we will see "consumer" versions. That is many years off. 

 Maybe.


3.23.2010

Media "Balance" Misreports Climate Change Impact


The mass media is blowing their responsibility to accurate reporting. (Image Source)

"Reporters need to learn that, if they wish to discuss 'both sides' of the climate issue, the scientifically legitimate "other side" is that, if anything, global climate disruption is likely to be significantly worse than has been suggested in scientific consensus estimates to date."
- Prof. William Freudenburg, University of Colorado at Boulder


A recent symposium on "Understanding Climate Change" at the 2010 Annual Meeting of the American Association for the Advancement of Science (AAAS)" discussed how even though the evidence for anthropogenic climate change (AAC) has become stronger and has scientific consensus, it remains controversial among the public and policy makers. The symposium highlighted the influence of the fossil fuel industry, conservative think tanks and others.

Presentations from the speakers can be found here

Further discussion and more links to informative background can be found here.


Climate Change Impact Worse Than You Thought

Here is an outstanding, up to date summary of the best scientific understanding of the climate with links to more in-depth discussion and source publications. The author, Dr. Joseph Romm is a physicist and Fellow of the American Association for the Advancement of Science. His blog Climate Progress is one of the best sources for climate change information.
"In 2009, the scientific literature caught up with what top climate scientists have been saying privately for a few years now: Many of the predicted impacts of human-caused climate change are occurring much faster than anybody expected — particularly ice melt, everywhere you look on the planet. If we stay anywhere near our current emissions path, we are facing incalculable catastrophes by century’s end, including rapid sea level rise, massive wildfires, widespread Dust-Bowlification, large oceanic dead zones, and 9F warming — much of which could be all but irreversible for centuries. And that’s not the worst-case scenario! The consequences for human health and well being would be extreme."

The following article in the same blog is also the best summary I've found regarding the specific impacts of climate change we can expect, especially if little or no action is taken: An Introduction to Global Warming Impacts

3.02.2010

UAVs for Police Surveillance and Arrest

A helicopter UAV equipped with infra-red cameras has been used for the first time in the UK to make an arrest. Thermal imaging transmitted to the officer operating the UAV enabled police to find a suspected car thief hiding in bushes in a thick fog. The UAV was originally designed for military reconnaissance but has been used in the UK for two years for search and rescue and to "crack down on anti-social behaviour". The drones are near silent and can hover or "perch" for hours.


The UK already has more CCTV per capita than any other country, and now these mobile platforms are adding to its reputation as the "surveillance state".

Weaponizing Mozart

Bringing the fictional authoritarian police behavior of "A Clockwork Orange" to reality, classical music is being used increasingly in Great Britain as a tool for social control and a deterrent to "bad behaviour". A school district "subjects" badly behaving children to hours of Mozart in "special detention" isolations. Unsurprisingly, some of these youth now find classical music unbearable. Recorded classical music is blared through speakers at bus stops, outside stores, train stations and elsewhere to drive away loitering youth. Apparently it works. Detentions are down, graffiti is reduced, and naughty youth flee because classical music now is "repugnant" instead of providing an intellectual and emotional opportunity to experience one of humanity's greatest arts.


Read About it.

12.25.2009

The Science of Avatar

A professor of astrophysics blogging as "Copernicus" has graded Cameron on the science of the movie "AVATAR".

As a professor of physics who has worked on SETI (Search for Extraterrestrial Intelligence) programs and exoplanetology, Copernicus's detailed essay gives AVATAR a "B" for astrobiology (scale of the alien ecosystem), an A+ for astronomy (choice of star system and setting), an "A" for astrophysics (atmospheric features), but a marginal "fail" on physics due to the floating mountains. However, in an update Copernicus revises his physics grade with a explanation for the floating mountains based on the superconducting properties of "unobtainium", a mineral figuring prominently in the movie.

Additional sources on the science of AVATAR are the pandorapedia and A Confidential Report on the Biological and Social History of Pandora.

In other words, even a hardened scientist skeptic (like me) can find a lot to like in this movie!