Thursday, January 19, 2017

Artificial Intelligence in our World

The fact is that technology will change our world. It has, it is and it will continue. If this is somehow arguable, I fail to see it. We may discuss how, and how much, but it will change our world. Recently Bob Wilson wrote about Fukoku Mutual Life, a Japanese insurance company, that will replace claims adjusters with artificial intelligence. This kind of change has been predicted for some time, I have asked a few questions about the impact technology is likely to have on the working American. Change is upon us, and we will adapt or we will not. We will succeed or we will fail. 

Recently, artificial intelligence has been in the news in other contexts. Last July, Forbes magazine questioned How A.I. Is About To Disrupt Corporate Recruiting. The article starts with a simple premise, that "corporate recruiting is broken." The opposite conclusion might have been our first reaction. We have seen a proliferation of Internet-based  tools for building, submitting and prioritizing resumes and applications for work. There has been a migration to these tools, and a notable population of employers will no longer even accept a paper job application. It seems to some that recruiting has never been easier. 

But, Forbes contends that these tools have also created vast information dumps. The ease of submitting applications and resumes has led to larger and larger volumes of candidates in each application pool. It says that all of this massive response to an opening falls "into a black hole, never to be seen again." I have seen some evidence of this. Individuals seeking state employment can prepare a generic application and then ask the computer to submit it for any future job postings for which the applicant might be qualified. When we post an opening, we often receive hundreds of applications, some of equivocal relevance.

Forbes suggests that an artificial intelligence (AI) named Mya is the "hope on the horizon." The author found this AI "sensitive to my needs, responsive, inquisitive, direct." This tool is designed to interact with applicants, both posing and answering questions. The goal is for Mya to deal with specific concerns and narrow the focus of recruiters. Mya will do the screening, and the recruiters will purportedly spend their time more efficiently, focused on "interviewing and closing offers." Certainly, a candidate may have questions Mya cannot answer, but Mya will ask the recruiter, and funnel answers back to the applicant, learning in the process. Mya may, over time, need to ask fewer and fewer question of the recruiter. 

Mya is expected to review applications, and to eliminate unqualified candidates from the pool of potentials. Thus, through sorting and culling, Mya will save recruiters time. Through communication with applicants, Mya will enhance their knowledge base, learn from them, and further cull the applicant base. In the process, it will shield those hiring officials from a multitude of questions and concerns. Developers think that the communication and attention Mya brings to the process will enhance recruiter performance and their image with job applicants. 

More recently, the British Broadcasting Company (BBC) noted another artificial intelligence role in the hiring process in Speech analysis could now land you a promotion. It describes the use of voice analysis by computers. Applicants and current employees are directed to call a number and interact with an AI computer that asks seemingly innocuous questions about the caller. The computer analyzes how the questions are answered. Not the substance so much as the manner of response. 

Through analysis of the responses, candidates may be selected for interview or hire. Employees may be selected for training, evaluated for promotion, or assessed for competency. The conversation with the AI is recorded, and how the human responded is quantified, measured, and evaluated. The AI uses characteristics like "tone of voice, choice of words, sentence structure" and evaluates the human regarding "openness to change, enthusiasm, empathy." 


The platform was built by investing significant time building parallels. Human subjects were tested using a variety of personality tests and techniques, and then their voice and speech patterns were recorded, categorized and catalogued. From that research, speech trends and indicators were identified, and the AI's parameters set. The programmers claim that the system is currently sufficient for decision making about 50% of the time, and further human interaction is required in the remainder. But that percentage is expected to improve. AI, by its very design, is expected to learn and improve with time.  

Some psychologists say that people decide how they feel about others very rapidly. One claims that you have about 7 seconds to impress someone. But the AI is quicker still. It "sums up your character" in a fraction of a second, and produces a series of "charts and diagrams" to describe you and your attributes for the recruiter or manager. 

Proponents admire the efficiency of the process. They also point to the breadth and depth of analysis, and say that no human could perform such in-depth analysis. They claim that "your voice reveals volumes about you," and that the AI can "decode you," breaking your attributes and proclivities into "500,000 aspects of speech." And from this, the computer will know all, define you, and effectively channel you into the appropriate future. Oh, Brave New World

The proponents claim that this technology will empower managers and human resource efforts. They stress that AI is "not intended to replace" human interaction. However, it will notably decrease the effort that leads to selecting which people are engaged face-to-face. It sounds like it may also alter the content of those face-to-face encounters. And human jobs will be lost in the process. 

While proponents claim this process will benefit applicants, the sentiment is not unanimous. Critics contend that these processes, at least initially, will disadvantage some, perhaps those who are not native-language speakers. They also contend that some will be able to manipulate speech patterns to enhance their prospects. Is it practical to believe that humans might actually manipulate computers? If a hacker can hijack a computer for data, or to impersonate, in other contexts, might they do so in this context? Maybe we will hire "voice coaches" to teach us how to speak effectively in order to gain an interview?

Universal acceptance of this process is not expected. But, proponents claim that it is the future, innovative, and cutting-edge. They suggest that companies that seek innovative thinkers may be the first to engage this technology. The mere decision by an applicants to embrace this process or not, may itself lead those companies to decisions about an applicant's suitability to work in a forward-thinking company. Those who eschew the innovation may be immediately and irrebutably seen as not viable contributors to such an innovative organization.  

If you doubt technology will change your world, you really need to get out more.