How do outliers affect a linear regression model?

They have no effect on the model

They always improve the model's accuracy

They can significantly skew the results and affect the slope and intercept

They only affect the y-intercept