When AI gets Really Scary

The Definition

AI is short of Artificial Intelligence. When you enable technology, whether a hardware, a software, a combination of both or a system or network of systems, to mimic responses or actions, that you'd otherwise need a human being to do with expended focus, attention and effort, you are developing Artificial Intelligence. In other words, something that is not intelligent or not supposed to be intelligent, operates, supposedly, intelligently, artificially so, being so programmed and designed to.

It is NOT About

If you are looking for affirmation of the Skynet scenario fantasy, you won't find it here. If you are looking for affirmation of paranoia or blame-game of AI taking away jobs, also, you won't find here. Those are not threats. They are misplaced misconceptions. AI does have its very real and credible downsides, in terms of dangers and threats. If you are interested in that, read on.

It is About

A painful and persistent problem with technology, is security. Especially, when connectivity is concerned. If something is digital, and it is not confined to standalone hardware or hardware with limited network connection, such as for example one computer being connected to another in the same room and nothing more, fundamentally, it can be hacked.

Technology can be used. It can also misused. And throughout technological history, it has been misused as much at least as it as been used for good, meaningful, constructive purposes. There is no credible reason to suggest or to believe, that any different should be expected with AI.

So what is Scary?

How hacking works, generally, through or across networks, at intelligent or more competent levels, is with the understanding of how codes of and for programs, software, applications, databases and so on, are configured and threaded. If a trained coder or programmer is able to figure or unravel how an email platform or portal or website or web accessible database codes are threaded and layered, he or she, can start systematically undoing those threads; for example, get into your inbox without your password by unraveling the construct of the email system.

The good news today is, that though anything and everything technological can pretty much be hacked, the process, if manually pursued, is very painstaking, tedious and time consuming. All cyber-security narratives aside, in the real world, the fact that hacking requires incredible patience, diligence, discipline and meticulousness, even with all kinds of malicious tools, applications and cyber-weapons that hackers may have at their disposal, forming thus a deterrence or a speedbreaker, that really creates the best defense there is today, keeping folks either from wanting to bother to hack into certain places or keeping them from being able to work through the layers of threads of code fast enough to successfully get access to something they should not have access to.

Your smartphone, tablet and laptop are not smarter or more intelligent than you. They don't have an IQ. You do. Yet they do things that you can't, at speeds you can't. And that is the precisely the point; speed. When something is digitized, it is automated, even if to varying degrees, and it works at a very different speed. A complex equation on a calculator is solved instantaneously, once you are done keying in everything. The same equation if approached with pen and paper may take an hour. The problem with AI, is that it enables automation to a very high and effective degree.

The rule of automation is simple. If it is quantitative or statistical, it can be automated. If is routine, patterned or sequenced, it can be automated. If there are limited parameters, it can be automated. All these things apply to coding. All these things also mean therefore, that technology can be designed to be artificially intelligent enough, to deconstruct and undo coding, in the form of Intelligent Hacking Agents.

What does it mean?

An Intelligent Hacking Agent, can mimic steps of human logic, except, much faster, operating autonomously once deployed, to study, configure, deconstruct and unravel codes, to alter, damage, shutdown or steal from whatever the set target. Just as a machine has been programmed to analyse, calculate and respond to possibilities, to beat a world champion at chess, an Intelligent Hacking Agent, can be programmed to do the same, with webs and layers of threads, of codes. The transformation of the hacking process here is, build something to hack, deploy it, go to sleep, and let that thing created to hack, go figure, hack, and do whatever it is, that is the purpose of the hack. And because, it is, digital, thus fast, it is harder to detect or protect against.

Like Intelligence Operatives, the human ones, Intelligent Hacking Agents may also be designed, to covertly embed themselves into files or software or systems, like spyware or viruses of today, to lie dormant, until they have analysed through configurations to attack, or until they are connected to a system they are deployed to hack. After all, there is a reason why the term 'Artificial Intelligence' was coined the way it was.

The most important implication here, is that the possibilities that an Intelligent Hacking Agent can be built to analyse, multiply astronomically, to speed up the hacking process. And that, changes the game, in hacking effectiveness. The number of things an Intelligent Hacking Agent can multitask on to do simultaneously, also multiplies astronomically

So why is it Scary?

Intelligent Hacking Agents probably won't be able to do what human hackers today can't. They will however, do what the human hackers of today don't or won't, because of the human limitations. And with greater efficiency and effectiveness, when speed, time, patience, effort and the number of things that can be done at a time, are no longer considerations. And working inside the systems rather than outside them remotely, Intelligent Hacking Agents will hold propensity for greater actual damage.

Who should be Scared?

Banks can see a lot of money transferred out, maliciously, by an Intelligent Hacking Agent systematically attacking all its systems on multiple front, simultaneously. They can otherwise be brought down will all records of what money they are holding, how much and where, to whom what is owed and so one. Even with sufficient physical backups, in this day and age, for many banks, that would be doomsday.

A thumb drive that inadvertently picks up a file with a hidden Intelligent Hacking Agent, brought into a weapons facility, by defense personnel one with access to triggers for nuclear missiles, can possibly trigger a nuclear catastrophe.

Energy companies can see their power grids hijacked by an Intelligent Hacking Agent spreading tentacles across them to bring them offline.

I can go on listing examples of who should be scared. There is an easier way to do this. Who should be scared? Everyone.

So we blame it on?

We will not create digital or mechanical life with AI. Not going to happen. If you wear tinfoil hats, don't add me on Facebook, I don't want to see your party pictures.

A gun is used to kill. It is the shooter, that takes the life. The same with tech. It is used or rather misused, for malice. The doer, is human.

Harish Shah is Singapore's first local born Professional Futurist and a Management Strategy Consultant. He runs Stratserv Consultancy. His areas of consulting and Keynote Topics include EmTech, Industry 4.0, HR, Digital Transformation, Marketing, Strategic Foresight, Systems Thinking and Organisational Future Proofing.