We hear a lot about AI agents these days, next-gen engines that are able to, in limited ways, act like humans and tackle tasks. But what about intelligent systems?
The intelligent system is something distinctly different from an agent, including in terms of the game theory that’s applied. Where an individual AI agent might compete with a human worker, for example, the intelligent system will seek to interact with dozens, or hundreds, or thousands of humans, and in some way, build capacity based on those interactions.
Defining The Intelligent System
You can get a concise definition of an intelligent system from this resource at GeeksforGeeks:
“Intelligent systems in artificial intelligence (AI) represent a broad class of systems equipped with algorithms that can perform tasks typically requiring human intelligence. These systems span various domains from robotics to data analysis, playing a pivotal role in driving innovation across industries. Here, we delve into the essence of intelligent systems, their core components, applications, and the future trajectory of this transformative technology.”
Authors provide the following list of prime aspects of intelligent systems:
- Reasoning
- Learning
- Perception
- Linguistic Intelligence
- Problem-Solving
You start to get a picture of how these systems might work, at least in theory. The intelligent system is working with us every day, learning at a global level, and applying its knowledge base to a wider world than the typical AI agent would have access to. That’s especially true in these early days of edge AI, where the agent can be installed on a non-connected, decentralized edge device.
Play Puzzles & Games on Forbes
Intelligent Systems in Gaming
Mike Ambinder has been in the gaming industry for 20 years. He’s an R&D partner at NEURAO, and has an evolved theory of how AI works in gaming and beyond.
In a recent TED Talk, Ambinder broke down some of these key concepts.
First, he contrasted games, with their interactivity, to other forms of digital experience: you listen to music, he pointed out, and you watch TV and movies, but you play games. That’s different.
By way of explanation, Ambinder broke things down into a linear process of a behavior that goes into a system and generates a response, and the cycle continues.
He also mentioned a term, “avoidances,” that represents, in his explication, the functions that are offered by a system.
What do you get with an intelligent system built for interaction?
I’ll put these in bullet points:
- Virtual environment
- Immediate and delayed feedback
- Dynamic altered and generative experience
- Continual user input
- Measurable behavior
- Systemic intention
A System with Goals
Ambinder further explained that an intelligent system has a systemic intention that is “goal-directed.” In other words, the system has its own greater purpose. That’s a hard concept to get your mind around, but in the age of semi-sentient AI, why not?
An intelligent system can also have incredible, far-reaching powers of surveillance.
“You can record everything,” Ambinder noted.
Intelligent Systems and Knowledge Generation Engines
Ambinder also described processes whereby an intelligent system can make sense of the data that it collects in what he framed as an “adaptive experience.”
“Instead of the player figuring out the game, the game can figure out the player,” he said.
That seems to be at the heart of this concept: that as you play, you, as the player, are not the only thinking party. The game will be looking to figure you out, getting more information about who you are, what you want, and how you act, as you play.
Use Cases for Intelligent Systems
Ambinder enumerated these key areas of use:
- Education
- Training
- Therapy
- Skill development
- Research tools
For example, he talked about treating PTSD, and how an intelligent system might be applied.
When you think about how these gaming ideas apply to AI, you start to see the ability to put everything in play, and have an intelligent system working on us, figuring us out, to some particular end. Presumably, it’s the owner of the system, probably a company or government agency, that’s going to benefit. Let’s make sure we talk about the rules for these evolved systems before we put them into implementation.
