Why the Top AI Firms of Tomorrow Are Today’s Masters of Diversity

We don’t just love to say it, we truly believe that Self-Awareness is the new gold. You can’t foster diversity without understanding your impact on others:

You should care about the topic of diversity for the sake of the validity and ultimately market value of your data product more than for any “soft” or cultural reason.

A lot has been written and said about why diversity is important as a means to attract talent and make better decisions, for examples. Whether you agree with these two things or not doesn’t matter because biases do impact the performance of the usefulness of data products, including AI. In this post, we advocate for the fact that diversity has to start with self awareness of identifying, asking for feedback on, and mitigating our own biases. We all have them. For the last several years, the classifications of these biases have been the subject of much behavioral economics research.

Cognitive bias is a limitation in objective thinking that is caused by the tendency for the human brain to perceive information through a filter of our own personal experience and preferences. The filtering process is called heuristics; it’s a coping mechanism that helps us to prioritize and process vast amounts of inputs. Its limitations can cause serious errors.

This video as an example of serious errors, provides an instructive example for the egregious inaccuracy as a result of bias in AI.

It likely isn’t totally possible to eliminate our brain’s predisposition to take shortcuts. Though understanding that multiple forms of cognitive bias exist can be useful when making decisions. Further, it can be helpful to take the gender and race stigma out of the conversation and topic of bias by focusing initially and perhaps exclusively on the definitions of the various types of cognitive biases based on the focus of the disciplines of neuroscience and behavioral economics. We find approaching the word “bias” this way makes people less defensive and emotional in ways that are counter productive. Plainly, we don’t believe the moral, ethical, or socially equitable reasons to pursue diversity will result in meaningful improvements. At Medigram we’re extremely practical and seek ways to drive productive behaviors; we don’t have an ax to grind on these difficult topics.

Our opinion on the top 10 Types of cognitive bias are:

  • Anchoring effect — the tendency for the brain to rely too much on the first instance of information it received when making decisions later on.

  • Availability bias — the tendency for the brain to conclude that a known instance is more representative of the whole than is actually the case.

  • Bandwagon effect — the tendency for the brain to conclude that something must be desirable because other people desire it.

  • Bias blind spot — the tendency for the brain to recognize another’s bias but not its own.

  • Clustering illusion — the tendency for the brain to want to see a pattern in what is actually a random sequence of numbers or events.

  • Confirmation bias — the tendency for the brain to value new information that supports existing ideas.

  • Group think — the tendency for the brain to place value on consensus.

  • Negativity bias — the tendency for the brain to subconsciously place more significance on negative events than positive ones. This bias probably evolved as a survival technique. Assuming the worst of a situation that turns out not to be dangerous is much safer than not expecting danger that turns out to be present.

  • Recency bias — the tendency for the brain to subconsciously place more value on the last information it received about a topic.

  • Survivorship bias — the tendency for the brain to focus on positive outcomes in favor of negative ones. A related phenomenon is the ostrich effect, in which people metaphorically bury their heads in the sand to avoid bad news. [1]

Cognitive bias and its impact on data management, analytics, and AI.

While data tools can help business executives make data-driven decisions, it is still up to the humans to select what data should be analyzed. Executives must understand that cognitive biases that occur when selecting data can cause digital tools used in predictive analytics, prescriptive analytics, and AI to generate false results that can cost money, lives, or elections.

Thankfully we can learn from history on the pitfalls of deploying and using predictive modeling without examining the data selected for analysis for cognitive biases. For one example, pollsters and forecasters predicted large margins of victory for Hillary Clinton in the 2016 U.S. presidential election. The culmination of many types of bias played a part in predictions that inaccurately forecasted Hillary Clinton would be elected president. Therefore, the reliance on weak polling data and flawed predictive models resulted in an unpredicted and unexpected outcome. This is not intended to shame the firms involved with the development and execution on the data strategies used; the point is to know that as a leader, our data products are impacted by biases and our jobs include our responsibility to mitigate these biases. This starts with radical self awareness of our own listed cognitive biases above. We need to understand how that relates to what and how data is selected, interpreted, and underpins any data product from analytics to AI. This is the mission critical business reason for making diversity of thought and experience real in our organizations.

[1] TechTarget Newsletter on Cognitive Bias

Eric Roth brings a wealth of relevant experience to Medigram as player coach and winner of numerous industry awards. This is in leading business operations and taking our culture game to the next level. As Head of Human Capital and Business Operations, Eric is excited about building a radically new kind of company in Medigram to win in the marketplace.

Together, with Medigram leadership, we have built the best team positioned to do what no one has been able to do. He is bringing his own recognized expertise to build and develop the strongest possible organization to deliver for physicians and patients. Eric is an expert in talent development, including management and leadership, having built proven programs in these areas. To win in our market, Eric enjoys also developing our extended team of collaborators.

As a business-oriented people leader, Eric Roth has driven rapid hiring and scaling for incredible 3x growth over two years. He is distinguished by his competitive drive to win. Further, he led people operations, systems, and compliance. Roth is committed to leveraging data and balancing it with common sense to drive successful employment and business outcomes.

Some highlights include having ensured target exit and funding events through leading people, operations, and organizational messaging; led the team to industry-wide recognition through awards by respected media.

Eric Roth is a transformational coach committed to developing people to their greatest potential. He has successfully led execution of several reorganizations and acquisitions, possesses a superb risk management record.

Most recently, Eric Roth was VP of Human Resources for a leading health care SaaS analytics company, Mede/Analytics.

Prior, Eric served as Vice President, Human Resources for Edelman, Inc., the world’s foremost global PR firm, and as Regional Manager, Human Resources for Paul Hastings, LLP’s US western region and Asia offices; a leading global professional services firm specializing in employment law.

Eric specializes in attracting and retaining top talent; charting, tracking, and measuring key performance metrics; and driving the success of an organization’s most important asset, its people.

Eric earned his M.S. degree in Human Resources Management from Golden Gate University with High Honors in 2003, his B.A. degree from San Jose State University in 1994, and is a certified Senior Professional in Human Resources.

Out of the office, Eric enjoys time with his family and traveling.


©2020 by Medigram.