“Artificial intelligence would be the ultimate version of Google. The ultimate search engine that would understand everything on the web. It would understand exactly what you wanted, and it would give you the right thing. We’re nowhere near doing that now. However, we can get incrementally closer to that, and that is basically what we work on.” —Larry Page
What is Artificial Intelligence
In 1950 Alan Turing defined the ‘Turing Test’ for the intelligence of a machine if it has the ability to exhibit intelligent behavior equivalent to, or indistinguishable from, that of a human. Turing proposed that a human evaluator would judge natural language conversations between a human and a machine designed to generate human-like responses. With the onset of chatbots, the perception has been created that these criteria have almost been fulfilled. A chatbot is a computer program which conducts conversations through text or through voice-to-text analysis. In 2016, the Washington Post reported that a Georgia Tech professor hired Jill Watson as a teaching assistant that throughout the semester answered questions online for students, relieving the professor’s overworked teaching staff. Only after the students’ final exams were handed in, was it revealed that Jill was, to the amazement of the students, an artificial intelligence bot. The question however begs whether Jill and all other artificial intelligence solutions marketed today are truly artificial intelligence.
Artificial Intelligence, according to the English Oxford Living Dictionary is: “The theory and development of computer systems able to perform tasks normally requiring human intelligence, such as visual perception, speech recognition, decision-making, and translation between languages”. Merriam-Webster defines is it as “the ability of a machine to imitate intelligent human behavior” and the Encyclopedia Britannica defines it as “the ability of a digital computer or computer-controlled robot to perform tasks commonly associated with intelligent beings”. Purest however insists that true Artificial Intelligence is a machine that exhibits self-awareness. Artificial Consciousness defines the aspects “that which would have to be synthesized were consciousness to be found in an engineered artifact” (Aleksander 1995). It is this version of the definition of Artificial Intelligence that Elon Musk warns about. A machine that possesses self-awareness might lead to self-preservation, and it is these aspects that drives the science fiction inspired fear. However, most people have a much more immanent and realistic fear and that is the fear of losing their jobs to a machine.
What is the difference between Machine Learning, Augmented Intelligence and Artificial Intelligence
Defining Artificial Intelligence is a challenge, as there are multiple definitions between the major players in this space and even between leading dictionaries. The challenge arose due to the underlying assumption of what Artificial Intelligence represents. Today marketing material widely uses Artificial Intelligence as a description of a specific solution, program or robot. The challenge in understanding what Artificial Intelligence is, becomes a lot simpler if you think of it as a field of study. And as with any field of study, there are more pure illustrations of the theory and sub-branches related to it.
As stated, the purest view of Artificial Intelligence is a machine that exhibits consciousness and the resultant self-awareness and potentially self-preservation. Achieving this level of Artificial Intelligence requires immense processing power and, in all likelihood, requires the onset of Quantum Computing to even be a possibility.
Quantum Computing uses quantum mechanics theories, such as superposition to derive more advanced processing power. The theory of superposition is realized into Quantum Computers through qubits. Classical computers use bits, 0 and 1. Quantum Computers augment the bits with a qubit that can be either a 0 or a 1 as a probabilistic model. Where classical computers can only be in a single state, either 0 or 1, Quantum Computers can be in 2n states – n being the number of qubits in the system.
Other definitions related to Artificial Intelligence are less stringent, but still require the machine to imitate intelligent human behavior.
Chatbots like Jill Watson, in general use question-answer pairing, rather than being able to imitate intelligent human behavior on a pure conversational basis. The complexity of the conversations that the chatbot can have depends on the complexity and extent of the tagging of information in the program. If you deploy a chatbot in a contact center environment which receives thousands of enquiries on a daily basis, each of which being tagged with an expert answer, then the level of conversations that chatbot can handle related to the subject matter at hand will become extensive over time. This however still does not imitate intelligent human behavior on a broader domain and the machine will make mistakes in a different way to humans. The selection of answers to the enquiries are still binary with the integrity and complexity of the answers being dependent on the data tagging.
To avoid the above complication, the term Augmented Intelligence is used instead, as it refers to the effective use of machines in augmenting human intelligence. This is a more accurate definition of the above-mentioned contact center chatbot, where the chatbot facilitates the transfer of information between the client and the agent. On enquiries that are fully programmed into the system, the Augmented Intelligence solution will seem like it can imitate human behavior, but for ‘unscripted’ responses the agent will step in. In a lot of cases the bot assists the agent internally to augment their understanding.
Machine Learning is a subset of the field of Artificial Intelligence which denotes machines that are able to ‘learn’ through statistical algorithms without being specifically programmed. Bots use machine learning to learn the appropriate response to an enquiry that is being tagged.
What exists today
In the purest definition, Artificial Intelligence does not exist today and based on what we know now, it is unlikely to exist until fully fledged Quantum Computers are being produced to enable the complex neural networks required to get close to the enhanced processing needed for a machine to develop consciousness.
Fully fledged Quantum Computers that are quantum in all regards are still very basic and only available at a very few advanced research labs. In fact, not even Microsoft has one yet. The couple of Quantum Computers in development by those few advanced research centres, use the principles of qubits, but from a hardware perspective are still considered to be classic computers in most regards and are being deployed mostly for research purposes and not general commercial use. (Giles, 2017)
Notwithstanding this, Artificial Intelligence is widely used in the industry, which should refer to the field of study, rather than the purist definition of an Artificial Intelligence solution.
Most solutions in play today that are categorized as Artificial or Augmented Intelligence, extensively use Machine Learning as the basis for their solution.
Augmented Intelligence is relevant where these solutions use advanced elements over and above machine learning, such as voice recognition, visual perception and complex decision-making.
True Quantum Computing might still be decades off, and although the threats foreseen by the musings of Musk’s of the world may not yet ring true, the risk of job displacement due to the rapid advancement of AI is very real.
Although Quantum Computing is not yet a scalable reality, we are increasingly incorporating and deepening our dependence on AI or at least Machine learning in our everyday lives, and the machines are getting smarter and more useful as a result of this, in a non-sentient way.
To think that ‘Alexa’ only came into being three years ago, how did we start our days without ‘her’? These days its almost impossible to imagine what life was like pre “Siri” or “Alexa.”
But the use of AI in our daily lives extends beyond the voice activated personal assistants and shapes our interaction with the world and our consumer behaviour.
With applications such as Netflix driving the high ‘on-demand’ personalized trends we are seeing by offering a predictive technology that makes personal recommendations to their customers based on their previous reactions and films choices. Amazons transactional AI is a perfect illustration of how using refined algorithms to predict purchases on the basis of online behaviour translates directly into financial gain, for Amazon at least. (Adams, 2017)
AI Examples in use today
Siri – Apple’s Personal assistant. Siri is a pseudo-intelligent digital personal assistant. She uses machine-learning technology to get smarter and better able to predict and understand our natural-language questions and requests.
Alexa – based on Amazon’s cloud-based voice service. Alexa can decipher speech from anywhere in a room. She can help you to find web-based information, shop, schedule appointments, start your day and power your smart home.
AI in Talent Management
Through our extensive research conducted for this book we can conclude that up to now the use of AI in Talent Management has predominantly been in the tedious, volume based, error prone areas such as recruitment, or payroll. “An IBM survey highlights that 46% of employees believed AI would transform their talent acquisition capability and 49% believed that it would transform their payroll and benefits administration.” (HR Technologist , 2017)
However, we believe that to truly benefit from the power of AI in Talent Management our focus needs to shift from the routine and mundane and into the space of hyper-personalisation and employee experience. This notion is supported by Jeanne Meister in her book The Future Workplace Experience, in which she is quoted as saying “An AI strategy for HR helps you create a much more personalized employee experience. You’re going to understand the real needs of your employees, and you better be able to deliver on those needs” (Jeanne C. Meister, 2016). Although theoretically there is much talk around the potential application of AI in Talent Management we are not seeing its application.
This is why we feel that it is important here to share with you a little more detail around the applications we have developed and successfully implemented. In order to demonstrate how AI can be used to enrich the employee experience and deliver on the promise of hyper -personalization.
For us, to deliver on this promise meant building Artificial Intelligence (AI) solutions in the Talent Management space, with a specific focus on Learning and Development and Career Mobility space. Hyper-personalised learning and career management to be exact.
Our AI solution known as 4th Talent makes an intelligent decision based on objective information taken from research over the last 60 years. This has been integrated into a decision-making hierarchy that measures the learning gap between the optimal performance profile and the individual’s profile. By integrating the theories of different fields of study into a non-contradictory decision-making hierarchy, we are able to identify reliable and valid constructs that underpin performance within multiple environments. For example, this could be multi-tasking and emotional management for a contact centre environment or strategic decision making and problem solving for an executive.
The key constructs are determined by an analysis of the tasks as well as the contextual corporate culture as they pertain to each role. We ensure validity and reliability of results by only leveraging research from the top global institutions, in other words institutions recognised for their peer review rigour.
Then we further corroborate our solution by integrating only those constructs found to be predictors of performance or retention through studies from a multitude of top institutions across various fields of study. Lastly, we understand and consider that every client’s environment is unique, meaning that we practically and statistically validate our customised solution for every client by testing and continuously improving our model utility (meaning the level of predictability of our model) within each client environment.
We have applied this in the following ways:
1. Learning and Development – 4thLearner™ Platform:
This platform enables Hyper-personalised Learning by providing an automated learning gap and proficiency level assessment and automated feedback report against a pre-determined set of skills. Based on the assessment outcomes different ‘baskets’ of curated content, matching the learner’s proficiency level, are made available to the learner for selection. This results in the learner being able to fully customise their learning path. Cyclical reassessment allows for a continuous personalised developmental loop. This encourages and reinforces a Self-directed, Agile, Digital, Learning culture. Learners are also able to rate and add to content creating a dynamic learner centric curatorship of a personalised developmental curriculum.
2. Career Mobility and Progression – 4thCareer™ Platform:
This AI solution automatically and scientifically tests, tracks and matches employees based on their competencies, characteristics, capabilities, faculties, experience, learnings and skills; with all potential future career path requirements measures. These assessments will determine the gap between the individual’s current profile and the optimal profile for the ideal future career path. This will generate automated, individualised career development plans designed to narrow the gap between the individual’s profile and the competencies, faculties, capabilities, skills and learnings required to maximise career performance. This individual career development can refer to generic interventions or it can refer to learning modules whose learning outcomes are matched to the skills requiring development.
3. Agile Workforce – 4thWorkforce™ Platform:
We use AI to scientifically and automatically assess, match, structure, deploy and quality control a project in a fully virtual consulting environment. We leverage advancements in scientific assessment techniques and sophisticated crowd sourcing technologies, to identify resources with the right competencies, capabilities, faculties, traits and skill sets, on demand. These resources are then scientifically structured into a project team and matched to the right templates, models and theoretical frameworks to optimise the potential outcomes of the project, taking into consideration the unique characteristics of the project, as well as the corporate cultural dynamics.
The above illustrates how AI can be successfully implemented right now and we are excited to share this with you in the hope that it inspires collaboration and spurs on innovative development and application in this space.
References
Aleksander, Igor (1995), Artificial Neuroconsciousness: An Update, IWANN, Archived from the original on 1997-03-02
Adams, R. (2017, January 10). Forbes. Retrieved from Forbs.com: https://www.forbes.com/sites/robertadams/2017/01/10/10-powerful-examples-of-artificial-intelligence-in-use-today/#3fb2e5d2420d
Giles, M. (2017, December 22). MIT Technology Review . Retrieved from MIT Technology Review : (https://www.technologyreview.com/s/609774/quantum-computers-barely-exist-heres-why-were-writing-languages-for-them-anyway/)
HR Technologist . (2017, April 28). How to Leverage AI for Talent Management. Retrieved from HR Technologist : https://www.hrtechnologist.com/articles/productivity-analysis-hr-analytics-tools/how-to-leverage-ai-for-talent-management/
Jeanne C. Meister, K. J. (2016). The Future Workplace Experience. -Europe: McGraw-Hill Education.