Hi!
This week we take a look at thin-slicing. Quick decisions, with limited information. Sometimes thin-slicing saves the world like in 1983. In 1983 when the world came very close to the brink of nuclear war. It could be averted because there were people who chose to ignore computer warnings, and made lightning decisions under extreme uncertainty.
The cold war was in fever pitch, with Americans under Reagan and Britain under Thatcher on the one hand, and an intelligence officer Yuri Andropov leading the Soviets on the other. In 1981, in a grand bureaucratic gesture, Andropov had created an institute which was running a program called Raketno-Yadernoe Napadenie (Nuclear Missile Attack) or RYAN.
The purpose of this computer model was to assign a numerical value to the probability of a missile attack on the Soviet Union. On September 26, 1983 in Serphukov 15 - an early warning secret command centre outside Moscow, a Lt. Colonel Stanislav Petrov was alarmed to see the computer's flash warning - five Minuteman Intercontinental Ballistic Missiles had been launched from an American base. Three weeks earlier a Korean airliner had been shot down minutes after entering the Soviet airspace, killing all onboard including a congressman from Georgia. The threat of an attack was thus palpable in the complex security establishment of the Soviet Union. Colonel Petrov was duty bound to inform the Kremlin of the impending attack, and the Kremlin in turn would have launched a full scale retaliation. Mr. Petrov received information from different sources within the command centre, and in perhaps what must have seemed like the longest five minutes of his life - decided it was not a credible threat and therefore did not inform his superiors in the Kremlin about it.
Colonel Petrov later explained that his gut decision was based on his distrust of the early warning system, and on the paucity of incoming missiles. Colonel Petrov trusted his training and intuition. It is clear what would have happened if he would have communicated the early warning to his superiors who would have turned to RYAN. The people manning the RYAN programme were people with a result in search of data. They were convinced in the autumn of 1983 that the Americans would attack. In our black box algorithmic society we need to have detailed examinations of how models work, what decision makers think, and where algorithms can go wrong. We need to combine rational decisions, with thin-slicing of decisions like the ones Malcolm Gladwell exemplifies in his book Blink.
“We are wiser than the computers,” Mr. Petrov said in a 2010 interview with the German magazine Der Spiegel. “We created them." (NYTimes, 2017). However, computer models are getting better all the time, a fascinating podcast from Data Skeptics on AI Decision Making and gets into an interesting territory of bounded rationality, which emphasises that decisions have to be good enough and not perfect. It has important ramifications for evolution of decision sciences & AI.
Want to check your decision making, try this, https://www.mindtools.com/pages/article/newTED_79.htm.
