I suppose expert systems or Knowledge Based Systems or AI might replace the human expertise in any field in the future. They are going to help human decisions by offering advice, so freeing human expertise. It allows organizations/companies/governments retain their expertise even as employees scarceness. No need of abundant credential holders or SMEs
It's just like doctors are taking the help of plethora of technical tools and computer systems assistance regardless of their smartness in their field of practice.
Interesting thought, albeit a bit vague, depending upon the definition of "future" - what is your horizon?
Historically, technologies have always replaced human for repetitive work, and AI will definitely accelerate the process. However, security is somewhat unique among I&T, thus I anticipated that CISSPs will be among the last to be 'replaced'.
Having said that, I have seen, during my career, certifications come and go. At one point, CCIE would open doors for 6 figure salary, but the developing SDN would in the near future reduce the need for network engineers significantly, especially the routine provisioning and fixing.
I have been advising, sometime unsolicited, my fellow 'younger' generations of I&T folks: keep learning or else.
Mouli @iluom ,
Expert and knowledge-based systems are a bit different from artificial intelligence (AI) systems. Building up an expert or knowledge-based system requires capturing the knowledge and expertise of humans in the field of endeavor. The idea is to capture their collective knowledge and also their collective logical and analytical processes, so that their collective capabilities can be applied in new, to the user, situations. Historically, physicians have been the only major filed consistently willing to share both their knowledge and their analytical processes with developers to be incorporated into the computer systems. Their positive participation is presumably due to the medical community tradition of consultation among the physicians. Practitioners in other fields have been reluctant to share their expertise, seeing the process as one of contributing to their own professional demise. When the bosses approach the experts with. "tell us all the details of how you work, so we can put it in this computer," the experts hear, "tell us everything about your job so we can lay you off as soon as the computer is smart enough."
AI systems are a bit different from the traditional expert systems because they rely on machine learning processes to discover and test new relationships and algorithms in their areas of effort. AI systems, once matured, can indeed assist practitioners, even experts, in solving both routine and advanced problems.
Your example of how physicians use such systems is actually the counter-example of your forecast that subject matter experts will no longer be needed. The physicians are the essential SMEs who must feed the situational data into the AI system and then understand and implement the recommendations of the AI system. In our security fields, we will continue to need SMEs who can take advantage of relevant AI systems to advance our work, too.
AI is good for the predictable world, looking at patterns and making "wise" decisions from there. One of the areas AI fails is when something unpredictable happens. A perfect example of this is criminals that run from the police. When a criminal decides to run they think they can either A) outrun their pursuer or B) outthink their pursuer; however when police units use K-9 units (dogs) the unpredictability of the dog foils the attackers. They know they can't outrun them, they can't hide their smell from them and they certainly can't "negotiate" or "smooth talk" their way out of an arrest/stop from them.
Expert systems have been just around the corner to replacing professionals since the late 70s. For a field that requires an understanding of business context and the ability to persuade senior staff I can't see the need for security SMEs disappearing any time soon. You could equally say all lawyers will be replaced by AI technologies, but again I can't see that happening due to the human dimension of working in organisations.
I was thinking any number of meta-information collectors in the BAD and NBAD realms as well. Either of these types of systems collect and tabulate far more information than any human can keep track and a good use of an expert system.
So agree with the above after I posted and thinking a bit more about how pervasive they already have become.
Nice ideas, and eventually repetitive activities such as hardening systems or initiating vulnerability scans and automatically remediating weak systems will occur, and in some ways is already happening within realms such as Red Hat and Cloud environments. However, remember AI and what mean is Augmented Intelligence, because we are a long way from actually achieving Artificial Intelligence capability. The biggest challenge at the present time, is actually the lack of computing power, because we are constrained by our current thinking and by the physical attributes of the materials we use. However, a deep rooted issue, which literally need ripping out and exposing is the inherent human condition - no matter what system we create, and how we are brought up and there is a vast danger of those inherent bias, and lack of objectivity ending up in the programming and education and learning capabilities of such systems. Understanding this bias, is vitally important, we may recognise it, but can we totally eliminate it from our developers and those who teach such brilliant systems we are inspired to create. There have been many futuristic films, which have been produced for entertainment factors, but there is an underlying tone, which needs to be recognised.
Is it the future we really want, and are we ready for it, and that we fully recognise the benefits and not the implications - which may extremely hard to go back, once we have stepped over the threshold.
As an example - I came into the office today, had an interaction with a colleague, and suddenly we have information we had not been informed about, but was vitally important for an engagement with a University. Would an Knowledge based system or expert or neural system have the same capabilities or interactions i.e. sharing information? Or would only be available, if you have the correct intellect, responsibilities and the right to know this knowledge before an important meeting that day? Would the Human and Augmented Intelligence in whatever form it manifests have the same capabilities with the next 20 to 50 years?