Skip links

Predictions All Over

Blog. Anything that can be predicted will be predicted. Such were the words in one of many  events in Davos during World Economic Forum 2019. Artificial intelligence, privacy and ethics and data for good use were among the hot topics in the Swiss mountain city.

“We are entering the age of prediction. Anything that can be predicted will be predicted. And we are getting better and better at accuracy.”

The words are Igo Tulchinsky’s (to the left on the picture). He is the founder, chairman and CEO of Worldquant, an investment firm with750 employees spread across 26 offices in 16 countries.

His company was hosting an event on predictions in the Hard Rock Cafe Hotel in Davos together with Jacquelline Fuller from and Robert kirkpatrick (second to the left), who’s director of the UN Global Pulse. According to Igor Tulchinsky the definition of prediction is; A forecast about the future based on the past. And the necessary level of accuracy depends on the area.

“In finance 51% accuracy is fine. In health, you want it higher,” he says.
And Robert Kirpatrick added to this: “When we talk starvation we talk weeks, when we talk malnutrition, we talk months. And in yet another situations it is about days.”

Davos from the top of Hard Rock Cafe Hotel

Robert Kirkpatrick has 75 data scientist sitting in New York city working with predictions. They can predict hunger by usage of the phone. Outbreaks of deceases by mobility info. Shipping data to detect rescue events and to help authorities be better prepared for refugees. They map hate speech on social media and hope to draw a global real time map of dicrimination around the world. They also want to build  a system on vulnerability to natural disasters.

According to Kirkpatrick privacy is fundamental in all data predictions, and as a UN institution working with data only for the good, it is not tempting to compromise the use of personal data to make money.

“We constantly focus on privacy,” he said and on the question of how to make sure data is anonymised he asked; “What is an acceptable risk of re-identification versus the risk of detecting a disease?”

He believes that telcos have some fine standards for anonymisation that other organisations and business could learn from.

Also read the Society of the Destination Machine of the Algoritmic God(s)