What’s the difference between AI and IoT anyway?
Recently I was asked the question, “What’s the difference between AI and IOT anyway?”. Also, “Because I’ve heard they are both kind of the same thing, aren’t they both about automation?”. And even, “Isn’t AI a movie?”. Well, I will expose what the difference between AI and IoT is, from what I’ve read and worked.
First I’d like to mention, even when I’m an Electronic Engineer I live in a software world. I work in a software development company, I’ve been in web apps development projects, as well as in the transition from those type of architectures to IoT architectures separated in micro-services. I know many people with different skills, and I have to say, the confusion between AI and IoT is also present in these kind of environments. I would say it’s very important for software developers to know all new stuff about uprising technologies and stacks, but even when we all want to we are so immersed in our working environment, we can get so in a confort zone, and technologies evolve so fast that we get behind so easy. So even though it’s just an excuse, it is a good one.
AI
As I mentioned, some might say AI is a movie from 2001 in which Jude Law was a Gigolo Robot, and yes, it is, but we are talking about technology and so the first thing to know about AI is it means: Artificial Intelligence.
Artificial Intelligence is not a technology itself, but a technological concept in the area of computer science by which the simulation of human intelligence done by machines is described and studied. The study of AI has been around for many years now and so the description of its purposes, activities, types or categories, composition, and examples is quite wide to explained fully in this entry., however, I will try to make a quick and concise but clear enough.
AI can be split in two, Narrow AI and General AI. Narrow AI is the one that includes an AI system designed to perform specific tasks such as Apple’s Siri, Virtual personal assistants and so many other we see nowadays. General AI or Strong AI on the other hand, is an AI system with generalized human cognitive abilities, it has to be adaptable, flexible and capable of learning so that when presented with an unfamiliar task, it has enough intelligence to find a solution. Now this type of AI does look like a description of Gigolo Joe from AI, or the Terminator.
Further than this we could mention concepts like Machine Learning, Deep Learning, Neural Networks, Supervised/Unsupervised/Reinforcement Learning, and about how AI solves problems of Perception, Knowledge, Reasoning, Planning and so much more, but as I mentioned before, that’s another entry.
IoT
Here I’ll be brief. I discussed the definition of IoT in a previous post so I’ll resume IoT as “a paradigm of networked technology development centered in the capability of physical and virtual objects to communicate. And so, IoT products, architectures, protocols, platforms and all layers must cover the ability of objects to communicate between them and to the world”.
Another thing to add to IoT is that it’s important the sensing and acting parts, and even the communication part of a IoT system should preferably be done without human interaction.
The following is an IoT skills chart from which you can get an idea what is involved in IoT projects and products. They all not need to be involved, but the bigger the stack more powerful it gets. Here you an also get an idea that AI is somehow involved there in the analytics side.
AI vs IoT
So by now the differences between concepts must be around your mind. IoT is about devices gathering info an communicating, AI is about processing info and learning from it. IoT can involve most of Engineering stack, AI is in the computers science world only.
Although they both share concerns on Laws, Ethics, Privacy and Security, their concerns lay on different layers of the stack, and they both raise different types of issues.
IoT needs physical phenomena and devices to exist and therefore create data. AI only needs data to feed data processing models to make the data meaningful.
AI and IoT
Now, should they be worked apart? Should they be thought as two different worlds? not necessarily. To me they are both complementary and a very powerful association of concepts.
IoT products or projects can stop as being monitoring systems, and that’s what we have most commonly seen around, but picture how powerful a system will become if using the data collected to learn how to behave with some physical devices as well. Possibilities are infinite.
AI systems only need data, but what about if they get their data not only from databases or already stored info, but also from continuously growing datasets fed by IoT systems, they will need eventually less “manual” work to be done and either less training steps. And so Narrow AI will give more options to General AI projects to appear, even ideas that have been around but doesn’t exist yet, such as Theory of Mind and Self-Awareness.
Conclusion
AI ad IoT are concepts that not surprisingly make a logical combination. Thinking that Smart-Cities for example will gather big amount of data, AI systems will enrich such types of projects in a way that will impact our daily life, and interaction with day to day devices and systems like Healthcare, Retail, Transportation among other.
Although both cocepts are quite very big to be handled in such a small entry, I hope the basic difference between them is a bit clearer, and again, this is just an appreciation, so it may change with more time, reading and experience.
Regards…
