What Is Assisted Intelligence?
Considered the most basic level of AI, assisted intelligence is primarily used as a means of automating simple processes and tasks by harnessing the combined power of Big Data, cloud and data science to aid in decision-making. Another benefit is that by performing more mundane tasks, assisted intelligence frees people up to perform more in-depth tasks. Requiring constant human input and intervention, assisted intelligence only works with clearly defined inputs and outputs. The main goal of assisted intelligence is improving things people and organizations are already doing — so, while the AI can alert a human about a situation, it leaves the final decision in the hands of end users. The exception would be those cases in which a predetermined action has been clearly defined.
What Is Augmented Intelligence?
The next level of AI is augmented intelligence, which focuses on the technology’s assistive role. This cognitive technology is designed to enhance, rather than replace, human intelligence. This “second-tier” AI is often what people consider when discussing the overall concept in general, with machine learning capabilities layered over existing systems to augment human capabilities.
Augmented intelligence allows organizations and people to do things they couldn’t otherwise do by supporting human decisions, not by simulating independent intelligence. Among the models included under this umbrella are machine learning, natural language processing, image recognition and neural networks.
The main difference between assisted and augmented intelligence is that augmented intelligence can combine existing data and information to suggest new solutions rather than simply identifying patterns and applying predetermined solutions. Thanks to deep learning capabilities and continuous training, augmented intelligence machines are able to make better and faster decisions than humans, which can be especially helpful in time-sensitive applications.
What Is Autonomous Intelligence?
The most advanced form of AI is autonomous intelligence, in which processes are automated to generate the intelligence that allows machines, bots and systems to act on their own, independent of human intervention. Once considered mainly the stuff of science fiction, autonomous intelligence has become a reality. The thought is that, like human beings, AI needs autonomy to reach its full potential. While autonomous intelligence applications are growing, organizations are not yet — and may never be —ready to hand total control over to machines. With this in mind, AI should only be given autonomy within strict lines of accountability — a belief that is in no small part due to those aforementioned sci-fi portrayals.
Additionally, autonomous intelligence is not a good fit for all applications, particularly those where it is difficult to quantify the best outcome. In these situations, AI can serve as an automated adviser, with humans retaining the responsibility of accepting and implementing decisions made by the technology based on any more qualitative, intangible factors that must be considered.
Reactive Machines, Limited Memory, Theory of Mind, Self-Awareness
Of the four types of AI, reactive machines are the most basic. Rather than storing and learning from memories or using past experiences to determine future actions, reactive machines merely perceive occurrences in the world and react to them. An example would be IBM’s Deep Blue, which was able to defeat chess champion Garry Kasparov by observing and reacting to the position of the various pieces on the board.
Limited memory machines are a step above reactive machines in that they do retain data, but only for a certain period of time and without adding that data to their library of experiences for future use. Many self-driving cars use this type of AI, storing data like the relative speed and distance of other cars, speed limit, and other data that allows them to navigate roads.
At the moment, theory of mind is only theoretical in AI as researchers attempt to build technologies that are capable of imitating human thoughts, emotions, memories and mental models by forming representations about the world and about other entities that exist within it. For example, the hope is to build computers that can perceive human intelligence and how people’s emotions are impacted by events and their environment in an effort to better relate to humans.
Like theory of mind, self-aware machines are not yet a reality. There are those who believe this to be the ultimate end goal of AI, with machines operating as humans do with an eye on self-preservation, predicting their own wants and needs, and relating to others as equals. However, there is debate about whether a machine can become truly self-aware, like Skynet in the Terminator movies.
How to Use Assisted, Augmented and Autonomous Intelligence
Regardless of the specific type, AI has numerous applications for federal agencies. Perhaps the most beneficial of these falls under the heading of assisted intelligence, with the technology taking over simple tasks currently performed by humans. According to Bill Eggers, executive director of Deloitte’s Center for Government Insights, federal employees spend about 4.3 billion hours per year on a variety of tasks, including recording and documenting information, handling objects and much more. Deloitte estimates that currently available AI and robotic process automation could free up about 1.3 billion of those hours by automating “the more menial sorts of tasks that most people really don’t like doing anyway,” Eggers says.
“We now have the ability to free up all of that time for more high-value, human sorts of activities, and that’s gotten a lot of attention throughout the federal government,” he says. “That enables quantum leaps in productivity for different employees and departments.”
Data management is another area where federal agencies could see major advantages. The federal government is digitizing more than 235 million pages of records, with the hope of reaching 500 million by fiscal year 2024, Eggers says.
“You can just imagine the value of intelligent machines processing this vast trove of data. With connected sensors and the Internet of Things producing even more data, this is a real sea change in how governments operate,” he says. “There is not going to be a technology that’s going to have as big an impact as AI on the public sector over the next 10 years.”
Recognizing its potential to significantly improve processes and productivity, federal agencies are working with companies like Google and Microsoft to harness the full power of AI. One example is Google’s work with researchers from NASA’s Frontier Development Lab to help identify life beyond Earth using Google Cloud AutoML to identify patterns in massive data sets.
“Google CloudML’s resources helped researchers root out false positives, rapidly classify light curves and identify key variables they hadn’t noticed yet, allowing data jobs to run in seconds, and at 96 percent accuracy,” says Mike Daniels, vice president of global public sector for Google Cloud. “This eight-week session of AI-fueled rapid experimentation and iteration guided researchers in the search for exoplanets, where intelligent life may still be waiting to be found.”