When a doctor makes a diagnoses she is pretty accurate. When a machines makes a diagnoses it is not quite as accurate. But when a machine and human make a diagnoses together they are more accurate than the doctor alone.
This is the example Minna Mustakallio from Futurice used to in her opening keynote in our Co-operating with the machines -workshop. The take away was that we, humans and machines, have very different strengths and weaknesses. That means that instead of fearing machines taking our jobs or planet we should consider how we could work together.
In her speech she compared AI to Sheldon, the super-scientist character in the series Big Bang Theory. Just like Sheldon, AI can be very accurate, make precise estimations, and quickly understand logical connections. But in social context it might not be quite get people.
That’s why we need to two to tango. Human see large entities. Ask questions and see social structures easily. Machines go through large amounts of data, repeat tasks precisely and create predictions of different scenarios. In other words, we complement each other.
Cases from Finland and South Sudan
For our workshop two real-life cases were introduced. First, Matthias Wevelsiep from Finn Church Aid described the war in South Sudan. A conflict that originally was ignited by two parties has spread and become more complicated over the years.
Hyperinflation, soaring food prices, and attacks towards civilians has led to 1.9 million people being internally displaced, another 1.6 million has fled to be refuges in neighboring countries, and as much as 40% of the population is severely food insecure.
So, what is the need for machines here? In this kind of situations data tends to get unreliable. As it comes from many sources it can be both confusing and contradicting. To draw a clear picture of the situation and development one needs to go through data and connect the right dots in it.
The second case came from Finland. Jaakko Mustakallio from Ellun Kanat introduced a recent debate from Finnish politics. A governmental work group was set to make cuts on corporate subsidies that the government had found in their own studies to be insufficient.
The work group, however, ended their work without any results as they couldn’t reach a common compromise. The group was widely criticized for not being able to implement any cuts. They were compared to the large cuts made to social and health sectors, and the group was criticized for taking aboard parties that benefit from the subsidies.
The surprising similarity in these two extremely different cases is the quality of information available. While in South Sudan it’s difficult to find trustworthy source for information, in Finland the different sides (those who support cuts and those who object them) argue with only the information that supports their side of the story.
First drafts of the solutions
Three groups workshopped on the two cases to find possible seeds of solutions. The data was in the center and how it could be utilized and enhanced to create more constructive environment for discussion.
Idea for the workshop was to identify how human and machines together could offer solutions. In other words, trying to identify the role of machines as well as humans in the given solutions.
For the two cases data reliability was clearly the most discussed issue. For example, the group workshopping on the corporate subsidiaries case came up with an idea of a news service based on people’s collective knowledge. But whose knowledge is reliable?
The groups found that machines could collect data and summarize it, but in both cases it would be difficult for the machine to evaluate the quality and correctness of the information offered. Even if the machines could check individual facts, humans would still need to use their expertise to see how all the details connect, and if they compose a meaningful entity.
Continuing with AI workshop
The ideas brought up in this workshop will be continued on the 17th of May in workshop called “Teaching the machines”. In this workshop we investigate the potential of machine learning algorithms based on the same cases introduced this time.
You can sign up for the rest of the events in Join a Workshop -page.
More about the topic (in Finnish): Podcast: Koneiden ja ihmisten välinen yhteistyö