Product Management Ethics in A.I. by former Yammer Dir. of Product

Machine learning and artificial intelligence are getting bigger and bigger every year, and they will be the hottest aspect of Product Management in the future. What is artificial intelligence and why do we need to think about its product management ethics? When we build natural language understanding, and movement activated products what do we need to take into consideration?

Former Director of Product at Yammer explains.


Former Director of Product at Yammer

Drew Dillon has spent the past decade in various Product roles, most recently VP of Product, Engineering, and BD at AnyPerk/Fond. Among other things, Drew delivered countless mobile releases, new products, and machine learning solutions.

Prior to AnyPerk, Drew was employee 60 and Director of Product at Yammer, focusing on user growth, engagement, and metrics. He’s an avid tweeter and blogger and passionate about emerging technology and its impact on society. He holds a Bachelor’s Degree in Computer Science.


Ethics in Artificial Intelligence

Drew discussed the incentives of Product Management, such as, when and how they go wrong. From maximizing the crave-ability of food additives to notification addiction, Product Managers have a profound impact on society. In the not too distant future, a number of those Product decisions will be delivered by artificial intelligence.

In this recent talk, Drew talked about what data quality, biases, and potential impacts mean. He also brought up ethical lessons from the history of Product Management and how we can learn from them to build ethical artificial intelligence.

Product Management Ethics in A.I. by former Yammer Dir. of Product


Bullet points:

  • Ethics in artificial intelligence is about not misusing the bad data. 
  • Product Managers need to remember to create value beyond stakeholder value. (Check: Salt Sugar Fat by Michael Moss.)
  • “All data has an opinion. You can’t trust the opinions of data you didn’t collect.”
  • Nir Eyal: The Hook Model. 1. External trigger 2. Drives the user to some kind of action 3. Leads to a variable reward (positive/negative) 4. Whatever it takes to drive the user to a deep level investment. These four stages loop around. (Hooked by Nir Eyal.)
  • The products are pushing the people further apart from each other. For example, liberals don’t have many conservative friends on Facebook and vice versa. The same happens on Twitter.
  • Example: Microsoft created a Twitter bot (Tay A.I.) that was unsuccessful. The bot ended up breaking laws and offending a lot of people online.
  • Example: Movement activated hand soap machine that gives soap when you put a hand under it. The problem was that it only recognized white people’s hand.
  • The products above are examples of the negative implication of interface isolations (pushing someone to a different place than everybody else, and benefiting majority while hurting minority.)
  • “Product Managers need to design systems with a north star beyond personal preference and metric optimization.”
  • We don’t get to A/B test the reality; we need to optimize more. 

Enjoyed the article? You may like this too: