"Most people see the benefits of empathy as too obvious to require justification", Paul Bloom:

...then I add, “I’m against it.” This usually gets an uncomfortable laugh. This reaction surprised me at first, but I’ve come to realize that taking a position against empathy is like announcing that you hate kittens—a statement so outlandish it can only be a joke. And so I’ve learned to clarify, to explain that I am not against morality, compassion, kindness, love, being a good neighbor, doing the right thing, and making the world a better place. My claim is actually the opposite: if you want to be good and do good, empathy is a poor guide. The word “empathy” is used in many ways, but here I am adopting its most common meaning, which corresponds to what eighteenth-century philosophers such as Adam Smith called “sympathy.” )

... I have argued elsewhere that certain features of empathy make it a poor guide to social policy. Empathy is biased; we are more prone to feel empathy for attractive people and for those who look like us or share our ethnic or national background. And empathy is narrow; it connects us to particular individuals, real or imagined, but is insensitive to numerical differences and statistical data. As Mother Teresa put it, “If I look at the mass I will never act. If I look at the one, I will.” Laboratory studies find that we really do care more about the one than about the mass, so long as we have personal information about the one. In light of these features, our public decisions will be fairer and more moral once we put empathy aside. Our policies are improved when we appreciate that a hundred deaths are worse than one, even if we know the name of the one, and when we acknowledge that the life of someone in a faraway country is worth as much as the life a neighbor, even if our emotions pull us in a different direction. Without empathy, we are better able to grasp the importance of vaccinating children and responding to climate change. These acts impose costs on real people in the here and now for the sake of abstract future benefits, so tackling them may require overriding empathetic responses that favor the comfort and well being of individuals today. We can rethink humanitarian aid and the criminal justice system, choosing to draw on a reasoned, even counter-empathetic, analysis of moral obligation and likely consequences.

...Strong inclination toward empathy comes with costs. Individuals scoring high in unmitigated communion report asymmetrical relationships, where they support others but don’t get support themselves. They also are more prone to suffer depression and anxiety. Working from a different literature on “pathological altruism,” Barbara Oakley notes in Cold-Blooded Kindness (2011), “It’s surprising how many diseases and syndromes commonly seen in women seem to be related to women’s generally stronger empathy for and focus on others.”

...In Consequences of Compassion (2009) Charles Goodman notes the distinction in Buddhists texts between “sentimental compassion,” which corresponds to empathy, and “great compassion,” which involves love for others without empathetic attachment or distress. Sentimental compassion is to be avoided, as it “exhausts the bodhisattva.” Goodman defends great compassion, which is more distanced and reserved and can be sustained indefinitely. This distinction has some support in the collaborative work of Tania Singer, a psychologist and neuroscientist, and Matthieu Ricard, a Buddhist monk, meditation expert, and former scientist. In a series of studies using fMRI brain scanning, Ricard was asked to engage in various types of compassion meditation directed toward people who are suffering. To the surprise of the investigators, these meditative states did not activate parts of the brain that are normally activated by non-meditators when they think about others’ pain. Ricard described his meditative experience as “a warm positive state associated with a strong prosocial motivation.” He was then asked to put himself in an empathetic state and was scanned while doing so. Now the appropriate circuits associated with empathetic distress were activated. “The empathic sharing,” Ricard said, “very quickly became intolerable to me and I felt emotionally exhausted, very similar to being burned out.”

One sees a similar contrast in ongoing experiments led by Singer and her colleagues in which people are either given empathy training, which focuses on the capacity to experience the suffering of others, or compassion training, in which subjects are trained to respond to suffering with feelings of warmth and care. According to Singer’s results, among test subjects who underwent empathy training, “negative affect was increased in response to both people in distress and even to people in everyday life situations. . . . these findings underline the belief that engaging in empathic resonance is a highly aversive experience and, as such, can be a risk factor for burnout.” Compassion training—which doesn’t involve empathetic arousal to the perceived distress of others—was more effective, leading to both increased positive emotions and increased altruism.

...Even I, a skeptic, would imagine there is some substantive relationship between empathy and aggression, since presumably someone with a great deal of empathy would find it unpleasant to cause pain in others. But a recent review summarizing data from all available studies of the relationship between empathy and aggression reaches a different conclusion. The authors of “The (non)relation between empathy and aggression: Surprising results from a meta-analysis” report that only 1 percent of the variation in aggression is accounted for by empathy...Baron-Cohen notes that people with Asperger syndrome and autism typically have low cognitive empathy—they struggle to understand the minds of others—and have low emotional empathy as well. (As with psychopaths, there is some controversy about whether they are incapable of empathy or choose not to deploy it.) Despite their empathy deficit, such people show no propensity for exploitation and violence. Indeed, they often have strong moral codes and are more likely to be victims of cruelty than perpetrators.

Peter Singer response:

Effective altruists typically donate a percentage of their income—usually at least 10 percent, and in some cases 50 percent or more—to charities that have been demonstrated to be highly effective. Some choose careers that will enable them to earn more not so that they can have more money, but so that they can donate more. Recent Princeton graduate Matt Wage, for example, was offered a place for postgraduate studies at the University of Oxford but instead went to Wall Street, where within a year he had earned enough to donate $100,000 to organizations helping people in extreme poverty. My admittedly impressionistic observation is that effective altruists are not especially empathetic—at least, not in the sense of emotional empathy. They do have what is sometimes called “cognitive empathy” or “perspective taking” capacity—that is, the ability to see what life is like for someone else...Unlike the majority of donors to charity, they are not prone to give to local charities, nor to particular children in developing countries who will write them thank-you letters. They do not give to causes that have touched them personally—“my wife/sister/mother died of breast cancer, so I donate to breast cancer research.” They direct the resources they have where they will do the most good. The result is that they are doing much more than most people to make the world a kinder and better place.