Jeremias Rößler (Roessler)

The end of banana software If the future of testing is AI, I will show you how. Various automation tools are available, but due to high creation and maintenance efforts for tests, GUI-Testing is still mainly a manual task. Meanwhile overall testing effort has risen to make up 30% of a typical software budget. Is crowd-testing the answer? What if we could have automated test cases be created automatically? Monkey Testing is not a new idea. But combined with a manually trainable AI and an innovative new testing approach (dubbed "difference testing"), we can now not only have the monkey search for technical bugs (i.e. crashes) but generate functional test cases that are optimized towards several goals and are even better than manually created ones. Visit the future of testing and see how AI can help us create better software!

Everybody is talking about the benefits of AI and where else to apply it. Few are talking about the dangers—except for (hopefully long-off) doomsday scenarios. This non-technical talk highlights the different dangers of AI, from intelligent autonomous weapons and public influencing to benevolent but misguided mighty ASIs, different problems are highlighted with their timescale and probability. A somewhat different talk.