Giving you personal and shared information on science, technology, technical & Investments in Nigeria

The Focus on Silicon Valley and Killer Robots Is Misplaced

The hotheaded development in deep learning over the last half-decade has brought transformed responsiveness to the topic of “killer robots.” Once principally an artifact of Hollywood’s imagination, independent lethal arms platforms are progressively becoming reality as both navigation and targeting technology improves with leaps and bounds.

Yet the outsized attention paid to Silicon Valley masks the fact that much of the most actionable work is occurring in the defense contracting and defense startup space outside the Valley.

Today’s focus on shaming Silicon Valley into halting the rise of killer robots is slightly different than efforts by US policymakers to halt encryption worldwide by limiting its accessibility and security within the US. Like encryption, AI is not a single secret algorithm known only to a handful of firms in the US and protected as a national security secret.

Advertisement

It is myriad of algorithms and approaches published openly in the literature and built all across the world. While Valley firms sit at the forefront of the AI revolution due to the enormous training datasets and talent pools they have, countless more advances come from universities and startups in many countries around the world. Most importantly, all of these advances find their way into the literature and turn out to be known to practitioners globally.

It is true that many technologies offered by Silicon Valley firms have dual purpose natures like facial recognition. It is also true that today’s correlative deep learning is largely domain-dependent, meaning that a facial recognition system trained on well-lit ground-level surveillance camera imagery is of slight apply in analyzing a video stream from a civilian drone looking down from feet above at night.

Deep learning technology today must still be largely purpose-built for a specific application, making it difficult to deploy an off-the-shelf facial recognition algorithm from a cloud vendor onto an off-the-shelf drone to magically turn it into a lethal weapon.

Advertisement

Instead, the reality is that while public interest has focused on Silicon Valley, the real advances in killer robots are coming from outside the Valley, by defense contractors and defense startups, with many key advances occurring outside the US in countries that have fewer qualms about turning to machines for an advantage in armed conflict.

Irrespective of what direction the US goes, other countries around the world will be progressively deploying lethal independent arms platforms.

Every week in Washington, DC and in capitals across the world there are myriad unsystematic presentations, workshops, roundtables, exhibitions and other demonstrations of independent warfare technologies.

While some of these presentations are by household name Silicon Valley firms, the massive mainstream of them is by firms whose names are few outside the defense orbit would ever recognize.

Some are even by universities receiving funds from the US Government or private firms to build key independent arms machinery.

Advertisement

In the previous year, we ourselves have observed innumerable demonstrations by both universities and defense firms and startups of civilian drones modified to autonomously navigate urban and rural environments in search of a targeted individual and then deliver a payload to them. Ostensibly such demonstrations typically describe themselves as delivering food or medical supplies to the targeted individual, however, the actual contents of the payload dropped from above are immaterial to the demonstrated technology: swapping a first aid kit for a grenade is all it would take to convert such a platform into a weapon.

A simple Google or Google Scholar search is all that is required to turn up many of these projects and some defense “presenter’s days” featuring these technologies are even open to the public. Given the sophistication of the applications that are publicly presented, one has to only imagine the systems already underway behind closed doors.

In some cases, these platforms are slightly more than off-the-shelf concepts like driverless cars or independent drones gave a lethal payload and modified to locate a specific target, while in other cases they are bespoke arms platforms built from the ground-up to utilize the full range of today’s deep learning capabilities in their work. The frequent reliance on off-the-shelf components like drones and inserted computers reminds us that almost everything has to be repurposed and raises the question of whether manufacturers have to do more to restrict the lethal apply of their systems.

The stress on Silicon Valley in efforts to halt killer robots is misplaced given that most of today’s advances are coming from outside the Valley and even outside of the US.

In many ways, killer robots are inevitable given the ease with which they have to be constructed from off-the-shelf tools and their low barrier to entry. Unlike nuclear arms which need enormous expertise and vast state-level resources, an off-the-shelf drone from the local hobby store has to be combined with an off-the-shelf microcontroller from the Web and run off-the-shelf open-source code to invent a do-it-yourself killer robot for just a few hundred dollars and an afternoon’s work.

Most importantly, it requires slight technical expertise to do so. In the end, killer robots, like encryption, are inevitable.

Advertisement
Advertisement
Spread the love
  •  
  •  
  •  
  •  
  •  
  •  
  •  
  •  
  •  
  •  
  •  
  •  
  •  
  •  
  •  
  •  
  •  
Advertisement

Leave a Reply

Your email address will not be published.

Advertisement
WP2Social Auto Publish Powered By : XYZScripts.com
error: Content is protected !!