girl with hand in robot's
PHOTO: Andy Kelly

Many people think of artificial intelligence (AI) as a completely automated process with no human input, but much of the data used by AI systems and many of the ways these systems are deployed are reliant on human input. In fact, despite fears that AI may replace human beings in the digital workplace, it is more likely that humans and machines will work together.

AI Augments Human Intelligence

People and machines are entering a new era of learning in which AI augments ordinary intelligence and helps people realize their full potential, according to Deepak Agarwal who heads machine learning and AI at LinkedIn.

Take the example of profile data, he said. At a fundamental level, almost all of LinkedIn’s member data is generated by members themselves. As a result, one company might have a job called “senior software engineer,” while at another company, the same role would have the title “lead developer.” Multiply this by millions of member profiles and it becomes clear that providing a good search experience for recruiters, where all of these varying job titles show up, can be a very challenging task.

Standardizing that data in a way that AI systems can understand is an important first step in creating a good search experience, and that standardization involves both human and machine efforts. 

“We have taxonomists who create taxonomies of titles and use machine learning models (LSTM models, other kinds of neural networks) that then suggest ways that titles are related. Understanding these relationships allows us to infer further skills for each member beyond what is listed on their profile,” Agarwal said.

Related Article: Getting Your Workers and Culture Ready for AI in the Workplace

AI Is an Actionable Tool

Even the term AI, which was a top buzzword over the past year will evolve from ‘hot topic’ into actionable tool, according to Ben Carlson, CEO and co-founder of Fizziology. As a part of this evolution, we’ll see brands add human intelligence to AI to make it more applicable and actionable to specific business goals. While AI is powerful for raw data analysis, human intelligence — applied after AI — can provide the context required for relevance.

He cites the example of a study the company carried out about the reaction to Colin Kaepernick as the latest face of Nike. There were more than 3 million mentions of the brand on Twitter during the first day of the campaign. While news of boycotts of the brand, burning and destroying merchandise drove headlines, those conversations only made up 3 percent of mentions the day of the announcement and less than 1 percent of mentions throughout the week after the announcement.

“While AI aggregated the mentions of Kaepernick from social media, it took a human to understand the true brand impact, contextualizing factors ranging from bots trying to sway the conversation to seasonal trends in social mentions,” he said. “To achieve marketing goals, AI and human intelligence will be a winning combination in 2019.”

Related Article: Workers Are Using AI at Home, Not the Workplace. Here's Why

How Humans Will Interface With AI

Pete Sena, CEO of digital marketing agency Digital Surgeons, argues that with the rapidly-evolving technological landscape, it is critical that we gain a better understanding of how humans are going to interface with AI, now and in the future.

At Digital Surgeons they define AI as augmented intelligence rather than artificial intelligence because, Sena said, they do not believe machine technology has reached that sentient level. Machines were first created to replicate a human brain — however, they are still lacking in emotional intelligence.

Ironically, creativity / human-centered activity will play starring roles in the future of companies using AI, human emotion and human logic must take center stage. It is crucial that the CIO understands the culture and capabilities of their team in terms of their ability to be self-learners who are continuously developing and evolving.

Emotional intelligence is just one component of the many considerations a CIO must weigh when evaluating his/her team. “At the highest level, CIOs need to be able to understand the capabilities of their staff in terms of being behavioral thinkers. I believe the the future is not with engineers and programmers, but rather in the hands of psychologists and anthropologists who have the ability to bring an emotional level to technology. One has to have an understanding of, or governance over, the machines,” he said.

In this respect, Natural Language Processing (NLP) is a critical skill to have as human-centered insights such as the development of decision trees will allow people to effectively visualize and structure data, which is paramount for the team’s success.

NLP is about turning language into actions and creating user and customer assessed teams who can understand both psychology and language. Humans need to be both the gatekeeper and translator between the customer and the machines. Smart creative people interfacing with machines is ultimately what will create competitive advantage.

In practical terms, Yu-Kai Ng CIO at TransPerfect said, this has pushed many companies to build AI to assist people in their daily lives. Large AI platforms like Alexa, Bixby, Cortana, Google Assistant and Siri are commonly used to play songs, listen to the news, check the time, set alarms and tell jokes. These platforms are evolving with all of the product development and data being fed to them to make them more useful, personalized and social.

“As AI matures we will continue to reap the benefits to a point where we do not recognize or attribute them to AI development,” he said. “Highly utilized navigation applications have had significant investment in AI development with their original purpose to take us from point A to point B to now making recommendations around those locations or taking us on an alternative route due to an accident that just occurred.

Related Article: Artificial Intelligence Moves Past the Novelty Stage to Practical Workplace Application

More Human Than Human

He added that in the translation industry, AI plays an important role in machine translation technology. AI is currently not as accurate as a human translator, but it can assist in speeding up the translation process with editing and review by a human translator to edit the content for accuracy.

In fact, according to Brian Jessup director of AI products at Yembo, humans will be able to be more human with AI. The current state of AI is that it can be trained to do the repetitive things that might be considered monotonous, but it struggles to create or adapt quickly to a situation it has not seen extensively. Humans will interface with AI, he said, through the typical means with which we currently interact with software, but the AI will be doing the heavy lifting of the repetitive tasks.

He cites the example of the household logistics industry. Here, he added, AI will be creating the inventory list from a video of the room, freeing the human agent to educate the customer and deal with the special one-off circumstances around the customer’s needs.

“Ultimately this will lead to humans spending more time using their empathy to serve more customers than they could before, while the AI relieved us of our boring tasks,” he said.

People are afraid automation will replace human positions. This is a reality, but if you look at the last 100 years of evolution in manufacturing automation, what you see is that automation replaces tedious, dangerous and low-skill jobs and makes way for a larger number of skilled positions that increase the value of the workforce rather than reducing it. In this current phase of this evolution, fear is based on the idea that AI will outthink people, Ben Schrauwen, CTO of Oqton said.

But, he concluded, automation being the most effective is when AI technology is combined with human decision-making, not replacing it. Humans are able to make creative and complex trade-off decisions in ways that AI simply cannot, despite the science fiction. Automation can allow skilled engineers to make better and more informed decisions by aggregating data gathered from the entire production environment in a way that people would never be able to. “I believe this technology will be incredibly powerful when people are using it to augment decision-making rather than expecting AI function on its own,” he said.