Game Theory Predictions of Paris Climate Talks
The upcoming United Nations Conference on Climate Change in Paris will be a negotiation to enact greenhouse emissions limits for every country. A recent study looked at how climate change negotiations take place using AI acting on game theory to conduct the negotiations. In each simulation each player simultaneously placed their bid of greenhouse gas emission reductions after trying to predict other players’ bids and offer what they predict is the lowest possible bid they could to still reach global emissions cuts of 50%. This procedure was repeated many times with varying factors including level of strategy of the players, number of players and more. A number of factors have been determined that could reduce the likelihood of a deal being reached in Paris. Not having a clear threshold about the point at which the effects of climate change will be irreversible makes it less likely that an agreement is reached which makes sense because if the players (each country’s negotiators) are unclear about this threshold, they will be unable to determine the payoffs for each bid they would make. The more players there are the less likely a simulation was to reach an agreement. The more varied the opening bids were the less likely an agreement was reached. Simulations with each player using a higher level of strategic reasoning actually resulted in a lower likelihood of a deal being reached. While each player is attempting to maximize his or her own country’s payoffs, this may derail negotiations and prevent a deal that addresses issues to the common good from being reached.
As a demonstration of the possibility for simulations of this kind to be successful in predicting outcomes of events, this article (http://foreignpolicy.com/2009/10/19/recipe-for-failure/) from before the 2009 climate change negotiations in Copenhagen predicted that conference’s inability to come to a strong or binding agreement on greenhouse gas emissions using game theory. Each country was faced with fewer benefits than costs for reaching the agreement. If they did not commit to the agreement, then they had no costs but still had most of the benefits due to free riding on the other countries who did commit to the deal.
Inaction over climate change and continuing greenhouse gas emissions is a kind of tragedy of the commons with the earth’s atmosphere as the commons, where when each country acts in its own self interest the common good will suffer. For these climate change negotiations to succeed will require an extreme amount of compromise by countries who each want to act in their own self interest.
http://www.popularmechanics.com/science/environment/a18060/ai-climate-treaty-game-theory/
http://www.nature.com/nclimate/journal/vaop/ncurrent/full/nclimate2849.html