Secret Service seeks sarcasm-spotting software

Of course Secret Service agents need a hand separating wisecracks from security threats.

Instead of searching for their sense of humour, however, the U.S. agency is looking for software to help them understand sarcastic remarks on social media.

It’s no secret that governments keep track of what’s happening on social media sites such as Twitter and Facebook, but Canadian experts are wary about whether or not sarcasm-detecting software will help security agencies keep citizens safe.

“I’m skeptical,” said Tamir Israel, a lawyer for Canadian Internet Policy and Public Interest Clinic at the University of Ottawa. “There’s a limit to what an algorithm can do, I think, is really what it comes down to.”

A tender document posted Monday shows the Secret Service wants to buy software that has the ability to “detect sarcasm.”

The Secret Service wants the software to also have the ability to identify influential figures on social media, analyze data streams, analyze old Twitter data and use heat maps. The work order specifies that the software should be compatible with the five-year-old Internet Explorer 8 browser.

Agency spokesman Ed Donovan said the request would allow the agency to create its own system for monitoring Twitter — both its own presence in social media and important issues that are trending on the social network. Detecting sarcasm is just a small feature of the effort, he said.

“Our objective is to automate our social media monitoring process,” Donovan said. “Twitter is what we analyze. This is real-time stream analysis. The ability to detect sarcasm and false positives is just one of 16 or 18 things we are looking at.”

The Secret Service currently uses the Twitter analysis program used by the Federal Emergency Management Agency, but needs its own, he said.

Using computers to find sarcasm may sound like science fiction, but it isn’t entirely out of the question, said Graeme Hirst, a professor of computer science at the University of Toronto.

“Tone of voice obviously gives you a lot. It’s hard to be sarcastic in text without the risk of it misfiring. Too many people will take it literally,” he explained.

Using software to detect sarcasm in speech would be much easier, Hirst said, but using similar approaches for text isn’t impossible.

“There’s quite a bit of work these days looking at many different aspects of emotion and affect and so on in text. And there’s no reason to exclude sarcasm from that.”

David Fraser, a Halifax lawyer who specializes in privacy law, said he has mixed feelings on government agencies looking for sarcasm on social media.

“It seems to make sense to use what technology you have available to you to kind of separate the wheat from the sheaf,” he said. “There are a whole lot of people who say a whole lot of silly things on social media, and if you’re legitimately using it to detect threats against the president and other protected people, you’d probably rather they filter out those who are just kidding, rather than go pounding on doors of people who are making some sort of snide, sarcastic comment.”

Twitter tends to be the platform where people go to complain, and users may not be able to fully express themselves in such a short format, meaning anyone who might stumble upon a tweet may not have the proper context for its meaning.

“If you have a nuanced complaint about the president, it’s hard to cram that into 140 characters,” Fraser said. “So somebody who’s emotional about that or upset, might abbreviate what they otherwise might say into something snippy or more curt, and perhaps more prone to misinterpretation.”

What government agencies do with the information they glean from social media is what concerns experts.

Tweets – sarcastic or otherwise – could be added to a profile without a person’s knowledge and cause that person real problems in the future, Israel said.

“Assumptions are made based on those snippets,” he said. “Then the next time they say something, they get flagged again. And eventually it builds to something that could lead to a more serious consequence.”

There have already been high profile cases of people running into trouble with the government agencies because of misunderstandings on social media.

A Twitter user was arrested in the Netherlands in April after tweeting what she claimed was a joke bomb threat to American Airlines.

In 2012, an Irish man and a British woman traveling together were taken into custody by Homeland Security agents at Los Angeles International Airport after the man tweeted that he planned to “destroy America.” He said “destroy” was slang for partying.

As more examples crop up, people become more likely to self-censor online, Israel said.

“Over time, we’re going to get more examples of these and as people become more aware of it, they do start to think twice about what they say and do online, in a medium where we’re trying to encourage people to just interact freely.”

That self-censorship is a huge concern for Sharon Polsky, president of the Privacy and Access Council of Canada.

“Our thoughts are under scrutiny,” Polsky said. “There are an awful lot of people in the world who just don’t get sarcasm, and I hope they’re not the ones who are designing the software.”