Remember meForgot password?
    Log in with Twitter

article imageOp-Ed: MIT — Rethinking the realities of Augmented Reality

By Paul Wallis     Oct 21, 2017 in Technology
Sydney - Between the teeming hordes of hype merchants and the very meagre actual products, augmented reality is taking some baby thinking steps into practical things. MIT is looking at making AR safe for users.
"Safe" means security, user safety, and dealing with the new range of unknowns that AR offers. If you haven't been very impressed with AR in the form of Pokémon Go or other descents into triviality, you may be thinking about safety from boredom and babbling idiots. The new reality, however, is potentially much more dangerous.
MIT is looking at practical security issues, like malicious software, physical risks to users, corruption of information, and all the other wonderful things we've come to expect from any kind of software. (The MIT article is much too short; it does address all relevant points, but really needed more space.) As MIT points out, one of the most critical issues with AR is its immersive nature. Users are almost literally glued to AR in ways other technologies simply cannot approach
Wearable, consumable, accessible, part of the Internet of Things, AR, like mobile before it, will be unavoidable soon enough.
It's not a pretty picture at the moment. Given the incredibly ineffectual nature of internet security, a multi-billion-dollar industry which has had absolutely no effect whatsoever on the spread of malware, AR does raise the risk factor to new levels.
One of the major problems is the likelihood of yet more dependence on technology. While overall technology has been delivering real value, it has also been delivering real vulnerabilities. A whole new class of risk has emerged out of the ideology of connecting the world. Connecting it to criminals and psychopaths probably wasn't such a good idea after all.
AR, however, is real time dependence. When walking down the street with your AR translator or other gizmo, exactly how good are your reflexes? Do you know a risk when you see one? You know how to get out of that risk?
Short answer, no. Acquired skills will help you navigate and work with AR over time, of course, but the learning curve could be tricky. The pity of it is that many of the new applications for AR actually are pretty good, and quite useful. That means you could be doing something important, when confronted with malware, software sabotage, etc.
MIT aren't the only ones thinking about augmented reality security, and the overall story so far is a list of potential problems. Risks to networks, in particular, are another much less obvious major threat. It is well-known that hackers exploit minor vulnerabilities in software, and complex software, naturally creates more minor vulnerabilities.
Personal risks aren't too encouraging, either. Many augmented reality applications use location tracking, which could simply tell someone where you are. Other possible risks include exploitation of biometrics, and other emerging and therefore risky data which is in constant use in real time.
A security fix for AR?
The simplest way to reduce risk from any technology is to be able to turn that technology off ASAP. If you are having any kind of problems with your AR, you should be able to ditch the application immediately, even if you're wearing it.
Protective software, of course, will be necessary, but how good is it likely to be? Perhaps an all-purpose "don't know what it is, won't run it" bit of coding would help, but would it work?
I can see a very likely scenario: SSL technology, at whatever degree of difficulty, that prevents new software and new apps from operating while the AR is functioning. It would be easy enough to introduce a series of protocols to prevent sudden strikes with "machine bureaucracy" telling something useful for a change.
AR and deep learning
Deep learning is usually associated with supercomputers and highly advanced computing processes. A probable future story is that deep learning will also be applied to consumer wearables and AR in general, simply because computer learning IS so useful.
How do you teach a computer to beware of malware? How do you teach much lower level AR software to do that? If you can do it, it might be a very useful achievement indeed. The likelihood, however, is that this will take some time, and risks will be present until deep learning can defend itself from our much-not-admired online parasites.
Let's just hope that an augmented reality world is a much smarter world. The world doesn't need "augmented" disasters.
This opinion article was written by an independent writer. The opinions and views expressed herein are those of the author and are not necessarily intended to reflect those of
More about augmented reality, Massachusetts Institute of Technology, augmented reality security, SSL protocols and augmented reality, Malware
Latest News
Top News