After Brian Walshe, should Google warn police of an impending murder?

Like other Internet companies, Google routinely preserves and analyzes every interaction with its users.David Gray/Bloomberg

How to dispose of a dead body. How to clean up blood. And can you throw away body parts? The shocking Google searches Cohasset resident Brian Walshe allegedly made around the time of his wife’s disappearance certainly look like damning evidence of a plan to commit murder.

So why didn’t Google do something about it?

Like other Internet companies, Google routinely preserves and analyzes every interaction with its users, in order to generate more accurate search results and more profits for its advertising business. So, could Google have alerted law enforcement that someone seemed to be using its service to prepare for homicide — and should it do so in the future?

In theory, search engines like Google could flag potentially scary search terms, just as social media companies like Facebook and Twitter flag threats of violence. But in practice, this just isn’t done, in part because of legal and privacy concerns.

Google did not respond to multiple queries from the Globe. But we aimed the same question at Bing, the search service run by Microsoft. A spokesperson replied, “We respect the privacy of our users in accordance with our privacy policy and do not monitor or act on specific searches they may run.“ In short, it seems you can ask Bing anything, no matter how grisly, without setting off any alarm bells.

To be sure, Bing retains records of every search, and those records are attached to the IP address of the device used to make the search. This IP data is deleted after six months, according to a privacy statement on Bing’s website. But until then, it’s possible for Bing to retroactively scour search data for evidence of wrongdoing and help police identify suspects. The company will provide this service to any law enforcement agent with a subpoena or search warrant.

In addition, police have developed a controversial practice called “keyword searching,” in which they ask companies like Google and Bing to hand over information about any users who have searched for certain terms during a certain period of time. Three teenagers on trial in Denver for setting a residential fire that killed five people in 2020 were identified when Google ran a keyword search to see if anyone had looked up the exact address of the house. This search produced the IP address of a device belonging to one of the suspects.

Brian Walshe listened to prosecutor Lynn Beland during his arraignment at Quincy District Court last week on a charge of murdering his wife, Ana Walshe. Craig F. Walker/Globe Staff

Civil libertarians argue that keyword searches violate the Fourth Amendment, because they don’t target specific individuals. But in November, a judge hearing the case said the search was lawful.

This is a far cry from the policies of social media companies like Facebook and Twitter, which keep an eye on postings in real time. They’re trying to ferret out messages with all manner of unsavory content, ranging from pornography to threats of violence.

Social media posts are aimed at individuals or groups of people. So if someone typed, “I want to kill you,” in a Facebook or Twitter message, it could well be a serious threat. That’s why social networks warn their users that hateful speech is forbidden, and why they use software and human reviewers to scour billions of messages every day, weeding out the scary ones and banning repeat offenders from further postings. Indeed, Facebook has a policy of notifying law enforcement immediately when members post messages “that involve the immediate risk of harm, suicide prevention and the recovery of missing children.”

But a query to a search engine is a request for information, directed to computers rather than people, so it’s not obviously a threat to anybody. People ask all kinds of questions for all kinds of reasons. Google Trends, a site that shows the relative popularity of search terms, reveals that over the past year, the search engine was routinely asked things like how to hide a dead body. The questioner might be a murderer, but could just as easily be a novelist doing research, a true crime buff, or even a police detective.

“Sometimes, people are just morbidly curious,” said Stephen Cody, a professor of criminal law at Suffolk University Law School.

Consider the case of New York City police officer Gilberto Valle. In 2012, Valle was arrested and eventually convicted of conspiracy to commit kidnapping, after his wife found Google searches and online chat files on his computer related to rape and cannibalism. But after Valle had served 21 months in prison, a judge tossed the conviction, because there was no evidence that Valle had ever harmed anyone. He’d only fantasized about it online. Prosecutors appealed, but a federal court upheld Valle’s acquittal in 2015.

If search engines were required to report suspicious behavior, there would be obvious privacy concerns.

Kevin Powers, director of the cybersecurity graduate program at Boston College, said Google could warn police about suspicious searches only by monitoring all user activity in real time. “Then we turn into China,” he said. “Everything you do is being monitored and surveilled.”

And even if Google alerted police when somebody searched for murder tips, Suffolk University’s Cody said such evidence would be insufficient to justify an arrest or even a search warrant.

“They could approach him and talk to him,” said Cody, but “I think it’s unlikely that any judge in the country would issue a search warrant based only on evidence of a Google search.”