Now What?!

Invitations Blog

October 12 - November 20, 2020

My Story, the Story AI Tells; Bias & Privacy

Updated: Nov 10

Presented by Alexia Georghiou & Moderated by Francois Knuchel


3 Tuesdays: Oct 27, Nov 3, Nov 10, 2020, 11:00am - 12:30pm EST


With the new world of AI, there are ethical considerations with implementation. We have been thrust quickly and deeply in this new world in 2020. We'll be exploring the implications of AI on our lives, our communities and on society as a whole, with a view of asking ourselves what are the ethical issues, what can we do about it and where do we take it from here. The session will be part presentation, with Alexia giving an overview of the AI landscape followed by participants working in smaller groups discussing the issues at hand.

Photo by Markus Winkler on Unsplash

Abstract

We all have a story to tell and have given over our privacy & data to a system that tracks, monitors and tells the story of us for corporations to sell us their products. The average person does not understand AI and its implications. I clearly explain how we are giving up our privacy, implications of bias using China as an example, and ethical considerations as we move forward. Marketing professionals benefit from AI and they need to understand how data mining to obtain consumer information could infringe on privacy rights.


Introduction

Artificial Intelligence (AI) is a computer science with machine and deep learning, that aims to solve cognitive problems commonly associated with human intelligence, enabling machines to “think like humans,” and perform tasks such as learning, problem-solving, reasoning, and language processing. Programs such as Amazon’s Mechanical Turk designate human beings to teach AI, by helping the computer to understand the mined data. In the year 2016, AI showed the ability to deep think by winning the 'Game of Go'. Google’s system studied the game, creating its own moves, astonishing onlookers. When it comes to our existing workforce, many hard skills are being automated with AI.



The Importance of Ethics & Values in the World of AI

In 2020 we have seen many happenings increasing our awareness of institutional racism and implicit bias as a society. The rate that AI is being implemented in our systems has not slowed nor paused during the worldwide pandemic. We are depending even more on automation to increase productivity among our teams to maintain profit margins. With AI, we can tell it anything we want to, and it will follow the programming. There are ethical implications that need to be considered as a priority as we enter this new age of AI; at the forefront being biased tendencies of humans feeding a flawed system. In general, the white middle-aged male is the majority demographic in corporate leadership, & the tech industry. Therefore, white males are the majority human beings teaching AI.

Arran Stewart, co-owner of Job.com gives us the insight with the power the humans in charge of AI have:

“AI’s race and gender biases are a result of who has the power in the backroom. Over 70% of all computer programmers are white males and despite our best attempts at neutrality, we were raised in a society that inherently devalues women and POC [people of color], teaching us both explicitly and implicitly that they are less capable than white men.”

73% of Amazon's leadership is male, and female leaders compose 32.6% at Facebook. An Amazon recruiting AI tool discriminated against women so they scrapped that tool. The data that was fed to AI included past performance of leadership which included a majority male pool of employees. The AI system therefore, was positively biased towards males in the candidate pool.

A recent government study on facial recognition bias found “demographic differentials” that can worsen accuracy based on age, gender and race.

Specific findings include:

· Asian and African American people were up to 100 times more likely to be misidentified than white men.

· High error rates for “one-to-one” searches of Asians, African Americans, Native Americans and Pacific Islanders.

· Women were more likely to be falsely identified than men, and the elderly and children were more likely to be misidentified than those in other age groups.

· Middle-aged white men generally benefited from the highest accuracy rates.


In response, Senator Ron Wyden (D-Ore.) said the findings showed how:

“algorithms often carry all the biases and failures of human employees, but with even less judgment.” In a statement, he added, “Any company or government that deploys new technology has a responsibility to scrutinize their product for bias and discrimination at least as thoroughly as they’d look for bugs in the software.”


In a research study published in Science Magazine April 2017,‘Even artificial intelligence can acquire biases against race and gender,’ researchers found that words like ‘female’ and ‘woman’ were more closely associated with arts and humanities occupations, while ‘male’ and ‘man’ were more closely associated with math and engineering occupations.


What this tells us is that AI will learn the gender biases in historic data and then propagate it into the automated decisions. There is widespread gender bias in the data sets used to train AI algorithms. There is a gender diversity crisis in the field of AI, which means that the ratio of people developing AI solutions is also skewed towards men.


What are some solutions for best hiring practices with the increased automation in our systems in addition to our human bias?


1. The Value of Soft Skills

“The key to overcoming this is to move away from traits such as experience and qualification (the AI can develop a bias for candidates coming from prestigious universities/companies) and focus on actual talent instead. This includes innovation, adaptability, and communication – “In other words, the things you can’t find on a resume.” Caitlin McGregor Co-founder and CEO at Plum

LinkedIn’s Global Talent Trends report found 92% of talent professionals reported that soft skills are equally or more important to hire for than hard skills. And 89% said that when a new hire doesn’t work out, it’s because they lack critical soft skills. These are skills that cannot easily be mastered by earning a degree or certificate. Soft skills can be assessed and taught through ongoing training & development.

Understanding the value in diversity & inclusion strategy with organizational culture is an outcome of soft not hard skills. If soft skills are difficult for human beings to master, and essential to successful employment, what are the implications with AI understanding this as an essential skill? We can begin with culture talk as the basis.


2. The Value of Diversity of Thought & Inclusion Strategies

Having laid a foundation for the need of diversity & inclusive strategies for AI and humans working together, we find great value in human diversity of thought. When we do not develop organizational culture intentionally, it creates itself. As people we feel, think, & speak thus behaving. We can hear and see the words and actions. What cannot be seen are the emotions. There are social norms in organizations telling us whether it is acceptable to express those emotions, or whether they should be held in and repressed. The best way to change and influence culture is through social norms.


There needs to be buy in from the team and teams follow leaders' behavior and example. When does culture begin to matter in the employment process? Culture development begins with the recruitment process and continues even when employees leave your organization. We begin with organizational values, grounded in virtues propelling us to create environments where teams thrive. Let’s integrate AI into culture change management strategies! We willingly feed AI our data on a daily basis. This includes the above components of culture: emotions, thoughts, social norms, actions, & virtues.


In a study atPortland State University, researchers found being a member of a marginalized group affects many workplace experiences, including engagement. These groups are often faced with overt and subtle discrimination and higher mental and emotional taxation both in and outside of work. It is therefore imperative for the success of our teams to foster inclusive strategies and an environment of belonging.


There is value in diversity of thought, as teams made of people with unique viewpoints help solve problems with optimal results. It is up to leaders to decide to seek and value diversity and seek talent from diverse groups. Equally important to this action, is to communicate the reason for seeking inclusion efforts. This includes connecting talent from underrepresented backgrounds with opportunity that those in the majority have unfair access to. Employees need to understand diversity and inclusion enriches the innovation, intelligence of the workforce with enriched ideas to flourish. An MIT study in 2014 found gender diversity at work increases productivity, as having diverse set of employees results in diverse skill sets. The study also found workers were most happy with teammates who were most like them. Even though the organization had strong social capital, this worsened work performance. In conclusion, the researchers found diverse skills led to increased bottom line earnings for the organization.



Meet Alexia

I motivate individuals & organizations to improve well-being with inclusion & belonging initiatives improving the overall culture. My passions include building resilience in healthy teams with empathetic conversations leading to innovation strategies, the ethics & bias with AI, and empowering youth to build diversity councils in schools.


Learn more about The Resilient Pathway HERE






  • Facebook
  • Twitter
  • LinkedIn Social Icon
  • YouTube