How to not kill people
06-19, 16:30–17:10 (Europe/Berlin), Kesselhaus

As AI grows, software manages more risks to humans. Moving fast and breaking things won't do. We will look at aviation to learn how successful risk management structures might look in software & AI.


With the rise of artificial intelligence, we give more control of our lives to software. We thereby introduce new risks, and the fatal Uber crash in 2018 is the first example of AI causing an accidental death. It will be up to us as software engineers to build systems safe and reliable enough to entrust with important decisions. Our culture, however, includes praising companies that move fast and break things (Facebook), celebrate principled confrontation (Uber), fake self-driving demonstrations (Tesla), and are right, a lot (Amazon). As an industry, we need to radically improve to meet the challenge, or more people will die.

In this presentation, we will look at aviation - the industry most successful at continuously improving safety - and attempt to learn. We will look at aviation safety principles, compare with similar practices in software engineering, and see how we can translate safety principles that have worked well in aviation to the software engineering domain.

See also: Slides (7.5 MB)

Lars Albertsson is the founder of Scling, a data engineering startup based in Stockholm. Scling provides data-factory-as-a-service - customer tailored data engineering, analytics, and data science. Lars is a frequent conference speaker on data engineering and data strategy. Before founding Scling, Lars has worked at Google, Spotify, Schibsted, and as an independent consultant, helping organisations create business value from data processing and machine learning.