Simon Smith debunks what people tend to call AI (Artificial Intelligence)

With respect, automation, machine learning, algorithms and programming that is based on human defined success factors is the definition of all of the above. Speech synthesis and recognition, OCR, gas sensors, camera and object mapping are all first programmed by a human.

Any further 'learning' is programmed by a success factor first programmed again by a 'human'. Machines run off instructions. This is a 'buzz' word that is misconceived, people have been achieving these results for decades, and there is nothing 'Artificial' about this method of design. It is intelligent, and some modern day applications are innovative, but the definition of 'Artificial' here goes only so far as to achieve simple tasks that humans naturally can do anyway. It is not intended to mean created by an alternate entity.

It is intended to mean exactly as defined.

DICTIONARY

adjective

"made or produced by human beings rather than occurring naturally, especially as a copy of something natural."

"her skin glowed in the artificial light"

Sadly there is no such thing as "consciousness" in computer science and logic does not cater for instructions made by non human created instruments.

It's just smart programming. Almost every application you would call AI, I or any competent programmer could create.

Especially in so-called cyber security solutions, I am seeing software claiming to solve a highly human caused and motivated issue, and it concerns me. I see many corporates paying hundreds of thousands of dollars for software that to me, I can see right through. It merely works as a blacklist of software, nothing more, nothing less. Nothing can detect polymorphic stenographic code that makes its way through a network and replicates itself randomly, and this has been the case from the beginning.

I put it to the Malware and Virus / software IDS providers out there to better prove their AI claims because it will not be solving the problems of any data breach, cybercrime or cyber security motivated act anytime soon.

I am the first to jump on any and every technology advancement I can see. I have created patents and been programming since a pre teen, even before my commercial days. Every example claiming to be AI, I (and others) could easily write an algorithm for.

I openly admit, many come down to brilliant ideas, like color matching in paint stores, voice recognition and Deepfake technology, but there is always a CPU, instructions, input and output.

What a computer system is actually doing when it is exercise "machine learning" traits, is looking for a pattern at such high speed that it gives the illusion a human could not do it. However, that argument is nonsensical because a human wrote the code. If you slowed everything right back down you will see reasonable human logic line by line.

There are no organs or parts of a computer system that 'can think for themselves' and often random is used to trick people into thinking a computer do just this. Random in itself is purely mathematical and only gives an illusion of different choices.

If one was to mention any form of "AI", I would be happy to reverse it into pseudocode and explain how the algorithm works.

It can be very smart, in some cases innovative. It can be automated, and can also be programmed to look through data for some form of 'success factor' which gives the illusion it has learnt something (as it cannot determine good from bad, or right from wrong) and this well known process is simply called machine learning.

Simon Smith, eCrime, Software & Social Media Expert Witness, Internationally Accredited Forensic Investigator, First Responder to real Cybercrimes CCISO CHFI CFE MCSD CLA MCSE MCSA MCT CEH CPT CAMS PRINCE2 PSM MPM COBIT5 ITIL AWS/CSA CSSBB