Thursday, November 29, 2018 at Heroku-Salesforce
We explored how machine learning and artificial intelligence are being implemented today. XLabs’ Travis Dirks led the discussion with our SME experts Jess Roeder, Anne Cocquyt, and Mars Hall on the 3 biggest causes of failed or mediocre AI including
- Not knowing, assembling or respecting the required circles of competence (The AI Triumvirate)
- Leaving validation criteria exclusively to the Tech Team and not establishing questions of business value such as 1) Is your AI even taking the right test? 2) How good of a grade does it need to get (based on the value/difficulty tradeoff)?
- Putting vast effort into determining if your AI is correct, and never asking if it is RIGHT (Lack of intent does not equal lack of bias.)
Then we separated into 4 ‘development teams’ where we collaborated, brainstormed and selected a product to focus on. Next, each team set its AI criteria, identified data sets and the right algorithm for their chosen ‘new product’. Last but not least, we had a sound off between teams to challenge utopia or dystopia of each team’s product at launch then at 20 years out.
- Utopians: How might your project lead to utopia (a place of ideal perfection)? How can you design to ensure it goes that way?
- Distopians: How might your project lead to an undesirable or frightening result? How can you design to ensure it does not go that way?
Not only was it great fun, we all learned a lot!
Meet our AI Experts:
Thank You Heroku-Salesforce.
Have a question or comment, let us know>>