AI and the rising expectations for software

One of the most interesting and powerful features of virtual assistants such as Apple Siri, Microsoft Cortana and the Google Assistant is their ability to become smarter.

Some of the functions of these assistants are taken for granted by users who are no longer impressed to get an accurate spoken weather forecast wherever they are; after all, how hard can it be for Siri to figure out location based on the iPhone GPS, fetch the forecast from a reliable website and apply text-to-speech subroutines to read conditions aloud?

The artificial intelligence involved in retrieving a weather forecast may not seem like much these days, but things are different when an iPhone user poses a command such as: “Hey Siri, call Chris’ office at 3:00.” First of all, Siri looks up in the contact list and sees entries for Christopher and Christy. The instruction of placing a phone call at 3:00 is ambiguous since it could be in the afternoon or in the very early morning hours.

Currently, Siri is programmed to establish disambiguation by means of questions. She could ask “Did you mean Christopher?” because it comes first in alphabetical order. There may be another question to clarify whether the call should be at 3:00 a.m. or 3:00 p.m. In the near future, however, Siri will no longer need to ask such questions thanks to cognitive computing, machine learning and natural language processing. Siri will know that Chris means Christopher because this is the contact that the user most frequently places voice calls to; furthermore, the virtual assistant will know that “office” means the landline at work and that the caller never makes phone calls after midnight. Siri will be able to understand, say “OK” and complete the task as expected.

Great AI Expectations

Until recently, the field of software development has largely consisted of programmers writing sets of instructions in a code that computers could understand. Programming languages and development platforms have evolved into code somewhat reminiscent of the human lexicon; however, the end result had always been to create binary or hexadecimal strings for the benefit of processors. In other words, developers were forced to learn the language of machines.

The time has come for machines to learn human language, and this is exactly what is taking place among all active virtual assistants. Microsoft Cortana and her competitors are constantly improving their voice interfaces through the process of machine learning, which derives its knowledge from massive natural language processing databases.

AI software evolution depends on the ability of computers to understand and interpret human language to enable natural speech communications. The fact that Siri can understand “Call Chris’ office at 3:00” is a major first step, but she is constantly learning from iPhone users because she is expected to vastly improve.

AI and the rising expectations for software

While virtual assistants are improving in terms of natural language processing, other software projects are getting better at predicting certain behaviors. The most common example is the autocomplete and suggest functions of the search bar in the Google Chrome browser. If an internet searcher in Fort Worth recently visited a few informational websites about scuba diving in Costa Rica, the next time she enters “flights from” on the bar, Google will suggest “Dallas to Costa Rica” or even “DFW to SJO.” In this case, the Google AI has analyzed browsing behaviors, location and search history to determine that the user may be researching a potential vacation.

In the enterprise world, the expectations for AI are even higher. Chinese electronics giant Huawei, for example, dedicates 40 percent of its workforce to research and development projects, which are heavily geared towards AI. This means that 72,000 tech workers could be working on smarter computing solutions at any given time.

One of the most visible AI projects undertaken by Huawei is currently being tested at a data center in Singapore. Instead of assigning computer scientists to figure out how to improve the efficiency of a data center that handles four terabytes of data per day, Huawei is operating machine learning constructs to speed up data processing and reduce the cost of energy consumption.

What is amazing about the fast pace of AI development is that influential figures such as Bill Gates, Stephen Hawking and Elon Musk are warning against the irresponsible or unethical use of AI in tasks such as predictive crime monitoring; to this effect, it is important to insist upon the responsible development of AI technologies.

Adam Richards

About Adam Richards

Adam Richards is a semi-retired business professional originally from Bangor, Maine. He spent the majority of his career in sales and marketing where he rose to the marketing lead of a Fortune 1000 company. He then moved on to helping people as a career counselor that specifically helped bring families to self-sufficiency through finding them rewarding careers. He has now returned to Bangor for his retirement and spends his free time writing. This blog will be about everything he learned throughout his career. He'll write on career, workplace, education and technology issues as well as on trends, changes, and advice for the Maine job market and its employers.