AUM logo CMYK-02-01
By Adrian • July 24, 2018

Tech Trends in A.I.

AI oneArtificial Intelligence (A.I.) has received an unprecedented boost in the past few years. Thanks to more funding and greater research incentives for both technology giants and smaller startups, exciting new development horizons are now within view. A bright future for A.I. in several industries as well as our personal lives seems guaranteed.

Much of this is due to the fact that more and more companies have acknowledged the need to embrace A.I. so as to stay ahead of the curve. Several large corporations with numerous employees are striving to become trendsetters in the field, despite the attendant challenges.

The visible impact of A.I.

What we used to relegate to the realm of science fiction is increasingly becoming reality. Indeed, thanks to A.I.’s versatility, we’re witnessing palpable progress in patient assistance, computational pharmaceuticals, and genetics, in addition to automation, robotization, and data management.

What’s more – as many experienced programmers, web designers, and tech consultants will attest – is that the birth of a new form of virtual marketplace allows for the sharing, selling, buying, and deployment of A.I. solutions. This has had a great impact on growth.

Data and intelligent machines

There is now a huge emphasis on data-driven models and the necessity for these models to be “smart.” A.I.’s machine-learning algorithms are quickly refining their ability to absorb and process data, reduce error margins, and facilitate seamless integration. This is often referred to as “deep learning.”

Due to the complexity of the computations necessary to sift through and make sense of mountains of information in a short period of time, the technology involved must be of a high caliber. Enter new graphics processing units (GPUs) with a redesigned architecture and a higher core count. This is of particular relevance to the business world, as the bigger companies place a high premium on mass data collection and collation.

A.I. on the phone, and in your home

AI twoA.I.’s advances occurred earlier than expected. For instance, experts assumed that no A.I. entity would beat a human player at the Japanese strategy board game “Go” before 2027. It turned out that the A.I. algorithm known as AlphaGo Zero succeeded in doing so 10 years earlier. Moreover, AlphaGo Zero was entirely self-taught, and developed game strategies unknown to human players.

When it comes to automated learning machines, Google made a splash last year with the launching of Google AutoML. This commercially available image-recognition tool can be configured without any experience in programming on the part of the user. What’s most impressive about it is that the intelligent machine is self-taught.

Following in the footsteps of Google AutoMLis NASNet, an A.I. mechanism that recognizes objects in videos being streamed. The combining of NASNet and AutoML has already resulted in models achieving results better than those brought about by human intervention.

Another form of A.I. making a difference in our daily lives is chatbots. Whether we’re making reservations or receiving bank account services, we’re often interacting over the phone with chatbots that register and process our requests. Indeed, in the span of just a few years, voice search has quickly leaped from being generally serviceable but sometimes unreliable to almost flawless. For instance, Microsoft recently carried out a test on its own voice recognition technology. Its error margin was a strikingly low 5.1 percent, which is a better score than that which professional human transcribers usually achieve.

Chatbots are also beginning to sound like us. Google Duplex, while still in its infancy, mimics human conversation over the phone and in the course of carrying out specific tasks. Long gone is the icy, monotonous, and robotic voice; in its stead, we have voices nearly indistinguishable from those that are human, replete with inflections and pauses. As it happens, this A.I. mechanism is having its levels of comprehension and response accuracy fine-tuned – so expect something close to perfect pretty soon!

Amazon’s “Alexa” is also worth mentioning here. These days, this chatbot is close to dominating the smart speaker market. Installing one in your home means that you will have a virtual assistant that’s forever on call; Alexa is always in standby mode, ready to carry out domestic tasks when you vocalize a request.

Both Amazon’s Alexa and Google’s Duplex function thanks to Natural Language Programming (NLP). Speech is something we humans take for granted, but one of the current challenges tech titans are taking upon themselves is getting their creations to interact with us in as organic and natural a way as possible. This means, among other advances, that the human speaker will no longer have to hew to a specific form of speech in order to be understood by machines.

Even our physical interaction with these machines is becoming, well, less physical. User Interface (UI) has gone from being grounded and tactile, such as with personal computers, to mobile and tactile, as is the case with most wireless smart devices, to aural and virtual.

IoT’s connectivity and centralization

AI threeDigital centralization is the next big thing. Having all kinds of technological devices in our lives, with each serving a different purpose, causes clutter, and necessitates the learning of countless modes of interaction. This is why, for the sake of convenience, a circuit-like connectivity between apps and devices, all controlled from a central hub, is so coveted. We’ve already mentioned chatbots and smart speakers, which serve as a communication point for users with this circuit. However, what is receiving more attention now is the Internet of Things (IoT).

We can define IoT as intelligent devices and apps working in unison in a collaborative environment. This network is connected to the Web and maneuvered by means of both wired and wireless communication channels. Regardless of the types of devices connected through IoT – be they health tech-related, smart vehicles, smart home appliances, or what have you – the deluge of data sent to a hub must be analyzed by smart machines capable of self-learning. This way, the circuit is optimized in real-time and anomalies resulting in failures are kept at bay.

Cybersecurity and questions of ethics

AI fourThere is a high likelihood that we’ll soon see one of the tech giants – Apple, Google, Microsoft, and Amazon – reach the trillion dollar valuation mark. While this may enable the company in question to wield unprecedented power and influence, it will probably also bring about increased competition, greater monitoring by regulators and lawmakers, and perhaps even a certain wariness on the part of many consumers.

As we saw with the Facebook and Cambridge Analytica scandal, the bigger the company, the higher the stakes when public trust is on the line. Will public trust in A.I. begin to wane? Difficult to say, but with smart technology becoming ubiquitous, and rising suspicions that certain smart home devices and A.I.-driven personal assistants are monitoring us on behalf of data-collection agencies, it’s possible.

If the public perception becomes that A.I. is overly intrusive, this could prompt regulators and lawmakers to effect a game-changing regulatory shift in order to protect users’ rights and place a cap on corporate power. Recently, we saw examples of something similar with the implementation of the General Data Protection Regulation (GDPR) in the EU, as well as the US’s net neutrality regulations. Yet too much regulation, of course, could end up stifling innovation.

Otherwise – as mentioned – the role of data is gaining greater prominence in the A.I. industry, and this has resulted in more devices and machines getting connected through IoT. A corresponding development, however, has been that hacking techniques are becoming more elaborate and effective. And with the overwhelming amount of data in circulation and the growing number of devices connected to each other, companies and individuals are a lot more vulnerable than before.

With this state of affairs in mind, much research is focused on curbing cybersecurity threats, and quite a few innovations are coming to the fore. For instance, machine learning takes a probabilistic, predictive approach to matters by analyzing the behavior of hackers and taking appropriate measures before protective firewalls and systems are breached. Another technological development in this arena is Zero Knowledge Proof, which uses mathematics to encrypt users’ privacy.

Meanwhile, something called CARTA (Continuous Adaptive Risk and Trust Assessment) is on the brink of making a breakthrough. Think of CARTA as constant data analytics. It examines data packages, downloaded files, and points and times of access – practically any activity that occurs in the network. While teaching itself to better identify and halt malicious attacks, it also tightens its network security.

As industry-leading companies grow, more players join the race, and greater progress is made in the way of A.I. tech, new questions surrounding ethics and responsibility are raised. For example, will A.I. entities eventually possess self-consciousness? If so, should they be granted rights?

A more immediate matter is the extent to which A.I., which has already taken over certain jobs and tasks, will replace the human workforce. It remains unlikely that A.I. will supplant humans entirely in the workplace, but this does not mean that major transformations will not occur. Already, many of us are having to adjust the manner in which we go about our jobs and further our careers.

Indeed, A.I., it increasingly appears, is here to stay. You’d do well to accustom yourself to its presence in the world at large, the workplace, and even your home. But do keep an eye on it – after all, chances are it’s doing the same to you!