An unethical optimization principle.

If an artificial intelligence aims to maximize risk-adjusted return, then under mild conditions it is disproportionately likely to pick an unethical strategy unless the objective function allows sufficiently for this risk. Even if the proportion of available unethical strategies is small, the probability of picking an unethical strategy can become large; indeed, unless returns are fat-tailed tends to unity as the strategy space becomes large. We define an unethical odds ratio, (capital upsilon), that allows us to calculate from , and we derive a simple formula for the limit of as the strategy space becomes large.

Nonlinear structures and thermodynamic instabilities in a one-dimensional lattice system.

The equilibrium states of the discrete Peyrard-Bishop Hamiltonian with one end fixed are computed exactly from the two-dimensional nonlinear Morse map. These exact nonlinear structures are interpreted as domain walls, interpolating between bound and unbound segments of the chain. Their free energy is calculated to leading order beyond the Gaussian approximation.