For Drone Researchers: “People are the problem”

chimp attacks drone

Missy Cummings is the director of the Humans and Autonomy Lab (aptly shortened to HAL; formerly known as the Humans and Automation Lab) at Duke University. A recent article by Todd C. Finkel in The Washington Post described her as “one of the nation’s top drone researchers”. The article was specifically concerned with current obstacles to using drones for delivering consumer goods as is being pursued companies like Amazon and Google. According to Finkel, Cummings “doesn’t doubt the technology” and “believes these autonomous machines already possess the ability to accurately and reliably do their jobs.” Finkel says that the technical issues regarding drone delivery have largely been solved.

But Cummings is concerned with the so-called “socio-technical issues”. These include kids throwing rocks, curious dogs, people with guns, etc. Indeed, searching “animals attacking drones” yields nearly half a million hits on Google and is becoming something of an online video genre. It is these sorts of obstacles that are purportedly delaying the Brave New World of drone delivery.

Or as The Washington Post succinctly puts it: “People are the problem.”

And isn’t this so often the case with technological progress? People are the problem. Or as one version of the Post headline put it “the enemy”.

The New York Times recently ran an op-ed in which investor Peter Thiel called for a vast expansion of nuclear power. He dismissed the accident at Chernobyl as “a direct result of both a faulty design and the operators’ incompetence”. Presumably the new designs Thiel is advocating for (and hoping to invest in) will be without fault and only put into the hands of inerrant operators.

Similar logic is employed to defend Google’s self-driving cars. The machine can do a better job than a human driver. It would therefore be best to eliminate the human element and turn operation of the vehicle over to the machine. Of course, the alternative that is not considered is that it may very well be the car that may need to be eliminated rather than the human.

The goal is to eliminate human error but to do that would require nothing short of eliminating humanity. Eliminating our animality. Assimilating us into a machine world or discarding us as refuse. While this is almost the explicit goal of transhumanists—those who hope to “transcend biology”—it is implicitly pursued by people such as Missy Cummings of Duke University. Cummings may not identify as a transhumanist, may not entertain the idea of living forever like Ray Kurzweil and Zoltan Istvan, may not pine to upload her brain to the internet, but she is doing the necessary work of overcoming the problem of people. Designing a world where our actions are without significance or impact.

In considering these “socio-technical” obstacles—the problem of human beings—Missy Cummings says she likes to consider how her eight year-old son would react to seeing a drone. She says “He’d like to throw rocks at it—because it’s there. It’s just human nature.”

While Cummings may see our rock-throwing nature as an obstacle to overcome, I see the impulses of her son as a source of hope. It’s an impulse that many of us share.

Pick up a rock.

Additional note:

drone researcherYou can learn more about Mary “Missy” Cummings at the website for the Humans and Autonomy Lab: 

The site lists her email address as:

The Humans and Anatomy Lab is located in Hudson Hall on the Duke University campus. (directions)

Lovebirds into Drones

Agapornis_roseicollis_-Marwell_Zoo-8aA study from three members of Stanford University’s Department of Mechanical Engineering found that lovebirds (Agapornis roseicollis) can rotate their heads up to 2,700 degrees in a single second. This allows the birds to maintain their gaze even while executing complicated, turn-on-a-dime, flight maneuvers. Furthermore, it was discovered that the birds are able to coordinate head motion and gaze direction with the flapping of their wings so as to minimize the time that their wings obscure their vision. While this is particularly useful for lovebirds who are fortunate enough to remain in a “dense and cluttered environment like a forest” where “proximity information about tree trunks and branches is essential” the abilities are, of course, also present in the birds who are confined in a university laboratory where PVC piping is the only available perching place.

Now this is all very interesting and it might, for some people in some small measure, increase the amount of respect conferred on the birds. But animal experimenters are quite skilled at tolerating the cognitive dissonance that must seemingly occur when their experiments confirm something remarkable about the the animals in their labs. Experimenters often justify initial experiments by talking about benefits that are purportedly conferred on animals but will almost inevitably call for even more experiments on even more animals regardless of their results. In short, if one didn’t already respect lovebirds it is difficult to understand how precisely quantifying their head rotation would prompt such respect.

Indeed, in the present case, the authors do not bother to suggest that a greater respect for animals may result from their experiment. The study received funding from the Human Frontier Science Program (grant RGP0003/2013) and the Office of Naval Research (grant N00014-10-1-0951). In the abstract for the HFSP grant, experimenters include half a sentence on one putative benefit for animals predicting that “results from these experiments will have implications for the design of bio-inspired aircraft and of bird-friendly, man-made structures such as wind farms” [emphasis added]. It is debatable how “bird-friendly” a wind farm might be but it should be noted that there is no mention of bird-friendly anything in the published study; in the published study we learn that their findings “can inspire more effective vision-based autopilots for drones.” The purpose of the Office of Naval Research grant has been summarized as follows: “to develop a bird-sized, unmanned aerial vehicle (UAV)…capable of navigating both urban and forest environments using vision-based control.” This explains the interest and the money provided by the military.

That humans need to humble themselves enough to learn from other animals has been a point that has been repeatedly made on this blog but experimenting on captive lovebirds so as to design better drones isn’t really what I had in mind.

Note: more information on David Lentink‘s lab at Stanford, where these experiments are conducted, can be found at: