AI: no need to be afraid

Santeri Kangas is the Chief Technology Officer at CUJO AI (opens in a new window). Santeri has 26 years’ experience in cybersecurity and cloud computing, with a stellar track record in building award-winning security software products for network operators such as AT&T, Orange, América Móvil, Telefónica, SingTel and BT. He shares five reasons why start-ups shouldn’t be afraid to embrace AI.

Tip #1

We live in a world with more data than the means to process it
We live in the world of Big Data, but data isn’t that useful without the means to process it and use it effectively for decision-making. That’s where I think AI becomes indispensable to modern companies. Any start-up that is delivering data-based services should pay attention to AI. I wouldn’t start doing anything with data without the right AI algorithm to process it.

Tip #2

AI will become ubiquitous in the near future
There was a time when we only found microprocessors in computers. Today all kinds of devices use microprocessors. I believe AI will follow a similar trajectory. Cloud computing has made it possible to store data remotely and, with the Internet of Things, all kinds of devices are now able to collect and exchange data. AI algorithms will therefore be present in more of these devices, helping process datasets not just in computers, but also in everyday devices such as home thermostats and key fobs.

As in the case of computers, technological developments are also driving down the cost of AI so it has become an inexpensive way to process large amounts of data, faster.

Tip #3

AI will help us make better decisions from raw data
AI is really effective when it comes to processing data quickly. It is able to process data 1,000 times faster than humans and it doesn’t get tired. This means that, so long as it has been configured properly, AI is far more reliable than human analysts when it comes to data crunching, and although humans are better than AI when it comes to making decisions, we’re handicapped when we have to use raw data to make those decisions.

Tip #4

Humans train and control AI
The common misconception is that AI has the capacity to develop itself. We are at least a couple of decades away from machines that have the cognitive capacity to develop themselves. Machines today are trained by humans. Yes, they can be trained badly or even for malicious purposes, but AI is not capable of misuse on its own. If there’s a risk of AI being used improperly, it is because the human mind behind it wanted it to function that way.

AI today is primarily trained using supervised learning. If it’s trained to do the right thing, it’s really effective at doing it. We’re also getting to a point where machines are able to spot the errors they make and learn from them, so there is little risk of AI committing a malicious act or a deliberate error without human involvement.

Tip #5

AI can be a great personal assistant
The advice I’d give start-ups is to explore the possibilities of AI. Even if you’re not looking to automate tasks or process large amounts of data, AI can be a great personal assistant and help you do your work better. I see AI-assisted decision-making as having a variety of applications. Take doctors, for instance. They will be able to serve more patients, and serve them better, by using AI to do the research for them. The possibilities are endless.

We shouldn’t be afraid of AI. AI, like computers, will improve our productivity. It will automate certain jobs, but it will not replace us. It will simply help us to do our jobs better.

Santeri Kangas © DR

The common misconception is that AI has the capacity to develop itself. We are at least a couple of decades away from machines that have the cognitive capacity to develop themselves.
Santeri Kangas