This first article “The Dangers of AI model drift: lessons to be learned from the case of Zillow Offers”, begins with the question “How much faith should companies place in AI making business decisions?” If companies were already holding off on AI, then Zillow’s experience will not help a bit. It makes a strong case for continuous monitoring of any AI system.
There are lots of good real-world examples when using a subset of data or a simple measure such as average can get you in trouble in the piece on “When Data Fails”. So true. Your AI team must understand the data at a deep level.
I also included a new blog from Scott Pringle that points out the amount of waste spent on planning in Scaled Agile and a little advice we give our customers on using on-prem for AI training and development. If you think you are not getting what you need from Agile, this is one place to start.
Enjoy these articles and the delightful weather through the weekend!
- The dangers of AI model drift: lessons to be learned from the case of Zillow Offers. How much faith should companies place in AI making business decisions? And how often should organizations ask this question of any particular model? The recent experience of Zillow Offers, an US online property marketplace, provides a cautionary tale about AI Model Risk Management. (The AI Journal)
- When Data Fails. In the early 1920s, car manufacturers had a big problem on their hands – engine knock. Engine knock is when fuel combusts inside your car’s motor in an uneven manner. This uneven combustion makes an annoying knocking sound and can cause permanent damage to your car’s engine. (Of Dollars and Data)
- The hidden cost of Agile. Agile rituals are filled with waste, so much so that sometimes we can barely see it! Removing artificial barriers and reducing the time spent on non-development tasks is paramount to improving productivity and reducing time to delivery. (SphereOI)
Just a final thought for this week. People talk about the challenges of democratizing AI by making it available to every organization. But with prices for using cloud resources for AI – especially with some of the large models such as BLOOM, GPT-3 (as examples), that is showstopper. Steep pricing for cloud services are forcing people to move to on-prem hardware.