Intelligent Systems Vs. Artificial Intelligence: How To Help Your Clients
If many of your clients don’t understand the difference between artificial intelligence (AI) and intelligent systems, you’re not alone. There’s a deeply rooted misconception about AI that isn’t going to clear up anytime soon.
AI has become a marketing buzzword and is being used interchangeably with computer algorithms that analyze data and produce a solution. For the majority of the population, this misunderstanding isn’t a problem.
However, when entrepreneurs and corporations start to pursue the development of their own intelligent projects, the lack of understanding can drive data scientists – and programmers – crazy.
Since AI seems to be what everyone wants, it appears that marketing messages are being constructed that tell people everything is AI when in fact, it’s not.
Instead of getting frustrated, here’s how you can help your clients understand what AI really is when they ask for an AI solution.
1. Explain that AI is marketed just like any other product
The most effective way to help your client understand when the solution they want isn’t actually AI is to explain that AI is marketed just like any other product. Everyone wants AI, so marketing messages tell people that everything they do is AI.
To a non-data scientist, it seems logical that if a computer predicts anything, it must be artificial intelligence. This isn’t true, and you’ll need to use familiar examples to override your clients’ beliefs.
2. Explain artificial intelligence to clients using familiar examples
Long before AI existed, machine learning was used to optimize actions based on patterns in data. Only when deep neural networks and reinforcement learning were applied did true AI emerge. It’s these two tools that gave rise to real AI like self-driving cars and machines that can read MRIs to determine if cancer is present.
Explain to your client that self-driving cars would never be allowed on the road if it weren’t for reinforcement learning and deep neural networks, and that each time a self-driving car encounters a new experience, it learns more, and that learning can be added to every other self-driving car’s programming. That’s how they were developed.
3. Distinguish intelligent systems from AI
According to Forbes.com, “artificial intelligence is the broader concept of machines being able to carry out tasks in a way that we would consider “smart.” Computers that are authentically utilizing AI technology are taught to think and learn for themselves, just like humans. This is in contrast to programming a computer with information and an algorithm to make decisions.
Explain to your client that intelligent systems work with patterns just like AI, except intelligent systems don’t have the deep neural networks that support self-learning. In today’s tech world, intelligent systems are being used by nearly everyone on a daily basis. Chances are, your client is familiar with at least one you can use as an example.
For instance, context-aware intrusion prevention systems (IPS) are now being used to increase security by scanning WAN and internet traffic. Cato Networks, for example, provides this context-aware protection with no capacity constraints. This is unusual for the industry but made possible by intelligent systems.
Cato’s intelligent system can detect suspicious traffic patterns and stop traffic from specific countries. It can use Active Directory to recognize user identity and perform true filetype inspection to prevent attacks.
4. Use the analogy: “a square is a rectangle, but a rectangle isn’t a square”
In school, you might remember learning that a square is a rectangle, but a rectangle isn’t a square. A rectangle is a quadrilateral with four right angles, and a square is a quadrilateral with four right angles and four sides of the same length. By this definition, a square is most definitely a rectangle.
By this same kind of logic, it can be explained that AI is being applied by using machine learning, but machine learning alone doesn’t constitute artificial intelligence. Machine learning is when computers can train themselves and adapt their own programming to the task. This is part of AI, but not AI itself.
Be careful if clients still don’t understand
Sometimes you need to let go of trying to educate a client on the technicalities they don’t understand. However, if you choose to proceed with a project for a client who insists on calling your intelligent solution AI, you need to be careful.
Avoid using the terms AI or artificial intelligence in your contracts to avoid being sued for not delivering on the project. Consult with your lawyer to create a clause that explains why your solution isn’t technically AI, and have the client initial that specific paragraph to indicate full understanding.
Some opinions expressed in this article may be those of a guest author and not necessarily Analytikus. Staff authors are listed https://www.datasciencecentral.com/profiles/blogs/advanced-analytics-platforms-big-changes-in-the-leaderboard