Artificial Intelligence (AI) developments have attracted much attention in industrial and financial circles. AI is being used commercially to “recognize spoken words, translate from one language to another, improve Internet search results” and to select financial investments (e.g. AIEQ, the AI Powered Equity ETF) that would normally require human judgement. Gradually, some workers will find their jobs have been taken over by AI systems and robots.
When AI replaces human workers, it’s impact is of concern. We saw this syndrome, when digital automation allowed employers to replace many clerical workers with computers. Fortunately, other jobs emerged and they absorbed some of the displaced workers.
AI may also displace workers in jobs that are considered skilled and resistant to computing advances, e.g. law clerks and medical coding clerks. Experts in the field feel that the number of jobs lost is expected to balance the number of jobs created by AI.
For an AI system to become proficient at its intended tasks, it usually must analyze a large database of examples, so it can learn “rules” that help it develop skills needed to answer the user’s questions. Some developer effort is devoted to preventing the acquisition of bad behaviors as the AI system makes progress toward the goals it was given. Instances of bad habits learned are probably a result of poorly stated goals and limits.
AI systems development requires highly skilled designers and programmers, but a new class of AI is reducing demand for those specialists. AI systems that develop other AI systems are in evolution. The specialists that guide these so-called Auto Machine Learning (Auto-ML) systems are a very rare breed and are in high demand.
Worldwide, only 10,000 specialists have the right background and skill for Auto-ML. They are particularly valuable because they can spawn many AI systems for commercial users. Companies such as Google and Microsoft have a strong interest in equipping their “cloud” products with AI suites that attract large customers. Most of the big names in Internet services (Facebook, Google, Microsoft, IBM and Amazon) are developing AI platforms for their customers to use.
Some AI developers are heading into the next sphere of computing – quantum computing (QC) – in large part because it is expected to vastly accelerate answers from AI systems. QC is an application of quantum mechanical phenomena, i.e., it’s physical. While QC can produce results far faster than conventional digital computers, it suffers from cost and occasional errors.
The scientific community has great faith that they will be able to tame QC’s cost and error infirmities, but it will take a decade to produce QC machines that can handle significant problems. To help pave the way to commercial QC, Microsoft made the Microsoft Cloud, Microsoft’s cloud-based environment for conventional programmers to become familiar with unavoidable, quirky QC concepts. IBM and Google have joined the field, but so far all the QCs available for customer use are simulated QC machines, not the real thing.
We can expect AI and QC to be in the high tech and employment headlines for many years. QC progress is closely related to encryption and national security. That may evoke some serious investments without accompanying headlines, if we are lucky.