Our knowledge to you.
One of the hottest trends in technology is autonomous agents and things. That nebulous-sounding term is actually quite precise – it’s a technology that takes advanced machine learning a step further, so that it can make complex decisions on its own, or autonomously. This is beyond simple automation, where something happens automatically according to hard-and-fast rules. Instead, autonomous agents and things make reasoned decisions based on multiple factors about the current situation – they choose actions designed to meet a certain goal without the involvement of people.
Examples include technologies like self-driving cars, advanced robotics, certain computer programs (including some viruses), or even something like a smart thermostat that senses when people are home and when they’re not, as well as other environmental changes, and adjusts accordingly – as opposed to one that is merely automated, running on a pre-programmed schedule.
This is an emerging technology, but we can see its evolution in technology most of us encounter every day. For example, virtual assistants like Siri (Apple), Cortana (Microsoft) and Now (Google) began as little more than voice recognition search functions, but are now much more sophisticated. In fact, in 2016 Apple announced that it is allowing third-party apps to access Siri, so that users will be able to ask Siri to accomplish tasks such as sending payment or searching images. Eventually the user experience of a smartphone will likely have an autonomous agent as the entire user interface, rather than a screen full of buttons for different applications.
Autonomous agents and things builds on the Internet of Things, in which devices are connected to the internet so that actionable data can be gathered. But the deluge of data provided by the IoT is becoming so overwhelming that it’s too much for humans to process. That’s where autonomous agents and things comes in — in the autonomous world, many technologies are interconnected and share data, and then act on it without the involvement of people. In fact, we’re now starting to refer to the Internet of Autonomous Things, or IoAT.
We’re not close to the point where an autonomous agent could take over the world, as has been depicted in numerous sci-fi movies (2001: A Space Odyssey, or Her). But there are some significant, albeit more pedestrian, challenges to be addressed.
Data security on the devices themselves is a significant problem, in that data can be easily recovered from decommissioned items such as smartphones – and people upgrade their phones at an extraordinarily rapid rate. And all kinds of IoT devices with capacity for storing and transmitting data are discarded frequently as well.
But virtual data security is an even more significant issue. As we’ve seen, the IoT is vulnerable to hacks and security breaches. Currently, the most pervasive problem is that devices are inadequately protected by passwords, leaving them open to be recruited into giant, impersonal botnets used in distributed denial of service (DDoS) attacks. But as we move toward the Internet of Automated Things, where decisions are being made based on data collected by these devices, the potential implications of such hacks – either to the devices themselves, or the cloud where the data is stored – could become more directed, and even more serious.
Furthermore, security issues – perceived as well as actual – might impact the growth of the technology in that they could cause people to distrust automated systems and things. We’ve already seen this effect with the IoT. It will be important for designers of automated consumer goods to learn from the mistakes of the IoT and effectively address security issues early in the technology’s evolution.
Another potential issue for automated consumer goods is that people might find them too complicated to use. If, for example, consumers pay extra to buy cutting-edge automated thermostats but get frustrated trying to program them, they’ll give up on those advanced features and just use the manual settings – and might think twice before choosing an automated product again. To avoid this, designers will need to pay special attention to the user experience as they roll out new products.
In the longer-term, liability will become more of an issue as systems become more and more autonomous – in other words, who will be held responsible if the system makes a decision that has harmful consequences? The manufacturer, or the owner of the system? It’s not difficult to imagine a scenario in which an autonomous system makes a decision that truly couldn’t be foreseen, especially as systems become more sophisticated. The regulatory framework will need to evolve along with the technology.
Computer programs are among the most well-developed applications of autonomous technology right now. For example, sophisticated supply chain management programs are capable of evaluating and reacting to needs such as ordering supplies, scheduling workers and so on without human involvement – going beyond simple automation.
Driverless technologies are already utilized in cars – for example, cars that can park themselves into tight spaces, or automatically brake when they get too close to another car or object. Evolution of truly driverless cars isn’t far behind — in fact, experts think this is possible by 2021. Ford, Nissan, Google, BMW, General Motors, and Daimler are just a few of the big names working toward this goal. Data security is of particular importance with this potential application, as the implications of hacking could be dangerous or life-threatening.
The world of autonomous agents and things is ever-changing. Keep up with your business’s advancing bandwidth demands with DQE’s secure fiber optic network services, where scalability is unlimited and customization is key.Tags: Artificial Intelligence, autonomous agents and things, botnets, Challenges, Current Applications, DDoS, Distributed Denial of Service, dqe, DQE Communications, Internet of Things, IoT, machine learning, security, technology trends