Today we mostly interface with the world through technology. We email, text, post information, consume information, get directions, find restaurants, create, and watch videos through our smartphone or tablet apps. This technology is becoming smarter and more capable.
In the 1980s and 1990s, software was purchased in a shrink-wrapped box. In the 2000s, software was downloaded to our computers. In the last ten years, software has evolved to become apps we use on smartphones and tablets. Today, many people use virtual personal assistants (VPAs), like Siri from Apple, Cortana from Microsoft, Alexa from Amazon, and Google Now. They allow us to do many things by verbally making a request.
Initially, VPAs were a novelty. They were good for marginally useful tasks, such as setting alarms, getting information, or initiating a phone call. VPAs are becoming increasingly more. They are doing more complex tasks.
We are on the verge of this autonomous-acting software changing how we live and work. Much of it is based on machine learning, where computers use algorithms to digest large amounts of data. In effect, they are learning without being explicitly programmed to do so. Machine-learning technologies have now advanced to the point where the computer can act by itself without intervention, and make more complex, autonomous decisions. This is the genesis of self-driving cars navigating our roads, autonomous robots doing manufacturing tasks, and advanced VPAs translating langauge in real time, making personal travel recommendations, automatically adjusting our homes, and even doing substantive work.
It seems we are moving at rapid pace towards a life where technology can make decisions on our behalf and provide smart suggestions for whatever we are trying to accomplish. A recent Gartner survey indicated that this practice of talking to our technology will become increasingly common over the next few years as VPAs become smarter and more capable. This technology is quickly advancing to the point where it can provide guidance and advice. That is where this technology gets even more interesting—and possibly frightening—in its potential.
- Beating User Fatigue: A recent Fortune magazine article discusses the phenomenon of app fatigue. There are too many apps—apps that don't get used, apps that consume space, apps that don't provide value. Users may not remember which apps they own. All of this may result in general user frustration or fatigue with using apps. Smart technology may be the answer. VPAs can work directly with a user and allow them to bypass apps. A recent Gartner survey indicates that within 4 years, 20% of smartphone interactions will rely on VPAs.
- Loss of Software Interface: By interacting verbally with smart software, the typical user interface or app disappears. The autonomous software becomes the interface. This "zero-touch" interface means the traditional software user interface goes away. Gartner predicts that by 2020 some 2 billion devices will be using a zero-touch interface.
- Ambient User Experience: As smart software advances and we adapt to using it, it will anticipate what we want to do with the discretion of an invisible butler. This invisible ease of use is a design methodology called the ambient user experience. The idea is that all the adjustments our devices make in our lives will feel natural and ambient.
- Very Personalized Technology: Smart software will understand more about their target user. It will not only have personal information about the user, but also possibly read their emotions, and learn their habits, interests, environment (light, temperature, noise), and location—all while anticipating what the user will do and acting one step ahead of them.
- Loss of Skills: Like any new technology, smart software makes it easier for users to access information and accomplish tasks. This may result in users losing important personal or professional skills related to how to navigate and retrieve information.
- Difficult to Trust: Any good relationship is built on trust. Users might not be comfortable trusting important work to intelligent agents. There will need to be some means to assure users that critical tasks can be accomplished by smart software as distrust for automated technology runs deep.
- Loss of Privacy: In order to do meaningful work for a user, smart software will need access to private information about that user. This opens up the possibility of privacy issues and debates on how such software will access and use confidential information.
- Potential Liability: Software that acts autonomously is much different than passive conventional software. Autonomous software controls itself and does not require human feedback to act. The software is responsible for its actions. On the other hand, conventional software is controlled by a user and its output can be directly tied to that user. Developers need to be more aware of potential liabilities when creating software that is autonomous.
- Detachment from Society: If the privacy and trust issues can be resolved, it is possible that some users may start to solely interact with smart software and communicate less with real people, causing a personal social detachment from society.
The above disadvantages are not new problems for anyone adopting new technology. We already face trust and privacy issues every time we buy something on a website, send an email, or access the internet. And, smartphone users are often accused of having detachment issues.
If One is Good are Many Better?
Smart autonomous software is no longer working alone. The next level for these programs is to work together. A multi-agent system (MAS) is when multiple agent programs interact with each other. Such systems can create solutions to problems that are logistically beyond what one autonomous software agent could do. Some examples include a city traffic system, military logistics management, traffic planning, complex financial portfolio management, and more. MASs can work across networks. They act as the interconnection for multiple older legacy systems to operate together, and they can manage complex project logistics on a global scale. Because many agents are working together, MASs are more reliable than a single software agent.
It appears that machine learning and big data are fueling the rapid growth of advantageous smart software and artificially-intelligent technology. The idea of an artificially-intelligent computer remains extremely popular in our imagination, but current smart software relies mainly on the less dramatic—and more pragmatic—analysis and interpretation of big data with machine learning. Today's smart software is not intelligent, but it does give a convincing appearance of being intelligent. However, this technology will progress. It will move beyond what just looks intelligent, and will actually become intelligent. It is possible that the rapid growth of autonomous systems will embody a fourth industrial revolution. Forward-thinking firms will start finding ways to use autonomous technology to leverage available data and transform how they do business.
- Barb Darrow. "How App Fatigue Is Taking a Toll on Smartphone Owners," Fortune. 16 August, 2016. http://fortune.com/2016/08/16/app-fatigue-is-taking-a-toll-on-smartphone-owners/
- —. "Why Smartphone Virtual Assistants will be Taking Over for Your Apps Soon," Fortune. 21 December, 2016. http://fortune.com/2016/12/21/virtual-assistant-usage-gartner/
- Tom Foremski. "Personal software agents versus the rise of the bots," ZDNet. 22 March, 2016. http://www.zdnet.com/article/personal-software-agents-and-the-rise-of-the-bots/
- "Gartner Identifies the Top 10 Strategic Technology Trends for 2016," Gartner Research. 6 October, 2015. http://www.gartner.com/newsroom/id/3143521
- "Gartner Says by 2019, 20 Percent of User Interactions with Smartphones Will Take Place via VPAs," Gartner Research. 21 December, 2016.
- Cecil Kleine. "DARPA wants to create Autonomous Agent Smith-like security software," The Next Web. 18 July, 2016. https://thenextweb.com/insider/2016/07/18/darpa-wants-create-autonomous-agent-smith-like-security-software/#.tnw_JGZwuvbj
- K.R. Sanjiv. "Investing in AI Offers More Rewards than Risks," Tech Crunch. 24 September, 2016.https://techcrunch.com/2016/09/24/investing-in-ai-offers-more-rewards-than-risks/
Back to Contents