The vitriol that swamped external affairs minister Sushma Swaraj\u2019s Twitter account following a controversy involving the grant of passports to an inter-faith couple is a dipstick of how dangerous a problem trolling has become. Swaraj, responding with signature maturity, flagged the particularly venomous tweets by \u2018liking\u2019 them\u2014one Twitter-user had advised her husband to beat \u201csense\u201d into her. She next conducted a Twitter poll, asking users of the social media platform, \u201cFriends: I have liked some tweets. This is happening for the last few days. Do you approve of such tweets ? Please RT\u201d. The fact that an overwhelming 43% thought trolling was OK, and by implication, that she was fair game, should jolt the nation\u2019s conscience. This means abuse, slander, rape\/death threats, doxing (publishing someone\u2019s private information\/details online with malicious intent) are now par for the course, and even legitimate reactions to political\/ideological differences. The Swaraj instance is, of course, the latest in an ever-growing list of examples of why Twitter and other social media platforms should police abuse proactively. The \u2018child-lifter\u2019 lynchings that WhatsApp forwards have fanned is another. Indeed, Twitter, in May, announced key changes to its algorithm that would allow it to judge the quality of tweets, to check if a tweet is adding to a conversation or otherwise. The idea, Twitter CEO Jack Dorsey stated, was to \u201ctake the burden off the person receiving abuse or mob-like behaviour\u201d. But, while Twitter et al work on the technological solutions to this, in the end, they are just amplifiers. What gets amplified is at the discretion of social media users. Differences of opinion, as Swaraj pointed out in a later tweet, are \u201cbut natural\u201d; however, criticism is most effective if it is done with decency. Sage words, but what about the audience? If trolling gets the kind of support it did in the minister\u2019s poll, then trolling is what will get normalised.