AI bets is a forum for the creation of adjudicatable predictions and bets about the future of AI. While it is easy to make a prediction about the future, this forum was created to help researchers craft predictions whose accuracy can be clearly and unambiguously judged when they come due. The bets will be documented on line, and regularly in this publication under the column name, AI Bookies.
Practitioners in many fields of science make leaps of progress when they advocate a side in high-profile “bets” as to what is possible or impossible in the next few years. Betting as a paradigm for stimulating advances in science has several interesting characteristics. First, betting requires two sides, and requires disagreeing parties to operationalize their differences as clear-cut claims with agreed-upon terms in order to determine if a bet has been won or lost (often this very process resolves disputes). By its nature, this approach requires claims that have some risk of failure, and thus avoids discussion of what is overly obvious. It also requires that someone care enough about a bet to work on it (whether or not that person or group is the bettor). Second, bets differ in important ways from competitive contests. A contest tends to drive everyone toward the same local maximum, with little novel contribution. In a contest, the goal is to be the best, not considering whether the best is actually valuable. A bet, on the other hand, can more explicitly carry implications that drive valuable discussion, e.g., by specifying the how in addition to the what.
A prediction is a single forcast of a future event or condition. A bet is a prediction having a challenger who is predicting a mutually exclusive outcome of the first.
The subject of the bet/prediction must be relevant to AI or a related field.
We want to encourage bets that promote or advance the field, and discourage bets like “when will be the next AI winter”, or “company xyz will be out of business by 2020” Bets are for reputation, not money. In some cases a bet will be the loser buys dinner, or a beer, etc., something cordial and friendly (to which the bookies themselves are typically invited). The period of the bet/prediction must be a stated length of time or until a particular event/situation occurs. In all cases, the conditions for time of adjudication must be clear at the onset. The Predictors and Bettors must provide an argument explaining why the subject of their prediction is important and why they think they will be proved right. All arguments will be in the public domain.
Individual people or small groups make bets/predictions using their real names. Large groups, formal organizations, and pseudonyms are not allowed.
Every bet must have an adjudicator. Adjudicators must be agreed upon by the bettors/predictors, and will be assigned if needed. The adjudicator must approve the operationalization of the bet. The outcome of predictions are usually decided through mutual agreement by the bettors; the adjudicator serves the role of being a neutral party who can resolve disagreements and the bettors agree to live with that resolution. Thus while the two bettors have the goal in specifying the bet to make their “side” as clear as possible, the adjudicator has the goal of ensuring the operalization is as clear as possible. There are no appeals.
Bets and predictions are always ultimately win/lose or true/false. There are no partials. This is important. A bet/prediction must be formed such that it can be adjudicated in this way. The data used to adjudicate a bet/prediction must be publicly available. There are no secrets.
Predictions may be promoted to a bet if a suitable challenger is found. Bettors must negotiate the exact terms of the bet, and this may result in changes to the original prediction. Bets/Predictions may not be revoked once final, unless through agreement by all parties.
The bet or prediction and the outcome will be made public.
Kurt Bollacker, Praveen Paritosh, Chris Welty