Show Hide image

How your Google searches can be used against you in court

Internet search histories are now commonly used as evidence in courts, but just how reliable are they?

On 7 October 2015, 21 days before the schoolboy Bailey Gwynne was stabbed to death at Cults Academy in Aberdeen, his 16-year-old killer made a simple seven word Google search.

It read: “How to get rid of someone annoying.”

This was by no means the sole reason the unnamed teenager responsible was sentenced to nine years in prison for culpable homicide, but it is a fact that many newspapers ran with in their headlines. The Sun called the Google search “chilling”. Anyone who has delved two pages too far into the results for “ricin” after a particularly riveting episode of Breaking Bad will feel the same – for entirely different reasons.

“It is now a common occurrence for computers to be analysed along with telephones and for all social media to be interrogated,” says Brian McConnachie QC, a criminal defence barrister. “It is almost commonplace for there to be reference to at least one of these in any serious trial.”

Still, unless the search in itself is illegal, it’s unlikely you’ll be arrested or imprisoned for your curious googlings about murder or human body disposal (note: feed to pigs). But if you’re already – rightly or wrongly – a suspect in a case, your search history can be considered very damning evidence.

“A one-time search can be evidentially significant and multiple searches are not required,” says McConnachie. “However, searching for a term on many occasions – like Lolita or such like – may make it easier to invite to the court to make an adverse inference.”

Bailey Gwynne’s killer didn’t just wonder how to shake off someone irritating, and also made a variety of other incriminating searches such as “Aberdeen stabbings death per 1,000?”, “illegal knives uk” and “difference between homicide and murder”. But his defence lawyer, Ian Duguid QC, claimed in court that the searches had been taken out of context and selectively chosen by the prosecution.

“It seems like no one has followed through in these searches to see what they are,” he said, referring to a video called “14-year-old Bronx student stabs bully to death outside school” that turned out to be a cartoon. Charles Bruce, the computer forensics analyst that examined the laptop, admitted that he only investigated the searches, not the links clicked on or their content.

If you don’t find it worrying that your internet history can be twisted and cherry-picked to defend or defame your moral character then congratulations, enjoy Farmville. But are you right to be worried? McConnachie notes that although a one-time search can be considered incriminating, timing is important.

“It's unlikely that searches from long ago would be used in respect of a recent crime as relevancy would be more difficult to establish,” says McConnachie, who defended Rachel Fee, a woman jailed for abusing and murdering her toddler, Liam. Before Liam’s death, Fee and her partner Nyomi had googled “can you die from a broken bone”, “morphine for children”, and “can wives b in prison together?”. During their trial, they pleaded guilty on a count of neglect.

“I have little doubt that it was the searches that made both accused accept knowledge of the fact that Liam had a broken leg and that in failing to obtain treatment they were therefore guilty of neglect,” says McConnachie. “Timing is also significant in relation to the date of the crime. Obviously googling ‘how to treat a broken leg’ is much more significant the day after someone potentially has their leg broken than six months before that date.”

There are also various defences that lawyers can use against this sort of evidence. McConachie notes that the accused can claim to simply have been curious about a matter. If they share a computer or a device, they can also claim they didn’t make the search.

“The main thing generally would be that you cannot tell from the search who searched it – although there may be passwords involved,” he says.

These are strong defences, should you ever find yourself framed for murder trying to explain the time you googled “how to destroy human teeth” to a jury. As long as you didn’t google it a hundred times the night before a body was found behind your house, you should be safe, right? Yes, provided the technology used to analyse your search history actually worked.

In July 2011, Casey Anthony was found not guilty of murdering her two-year-old daughter Caylee. Many members of the public were outraged, as Anthony’s partying personality and alleged inappropriate grief were considered evidence of her guilt. Others could not dismiss the fact – referenced extensively in her trial – that an investigation of Anthony’s home computer revealed she had searched the word “chloroform” 84 times.

Except she hadn’t. The software used by the witness John Dennis Bradley had been faulty, and he later discovered that the world “chloroform” had only been searched for once. The website Anthony then clicked on was about the use of the organic compound in the 19th century.

The case illustrates that not only can search histories be faulty, they can also be very prejudicial. If I told you that Justin Ross Harris searched how long it takes an animal to die in a hot car a month before his son died in the same way, would you assume his guilt? If you’re an infrequent googler, maybe. If you Google every question that pops into your head, maybe not.

If all this has scared you into jamming down the delete button in a panicked frenzy, stop. McConnachie notes that investigators can and do recover deleted searches. “Indeed the fact of it being deleted is possibly even more incriminating as it could be said to demonstrate that you have something to hide,” he says.

There are two lessons here. If you are a criminal: use a library computer. If you are just curious, researching a crime novel, or fact-checking your favourite TV drama: don’t be a criminal.

“If we intend to get involved in crime or if we have been involved in crime then beware your internet search history,” says McConnachie.

Amelia Tait is a technology and digital culture writer at the New Statesman.

Image: Getty
Show Hide image

Man makes $4bn in two days explaining Facebook to old people

Mark Zuckerberg's supposed blockbuster grilling by Congress was the bust it was always going to be, and he went home victorious largely by default.

On Tuesday a crowd gathered on social media for what promised to be a generation-defining moment, like the moon landing, or the OJ bronco chase. There was an air of tension. Mark Zuckerberg, founder of Facebook, was about to be dragged before the public and made to answer the Questions Of The People.

Many tuned in expecting a spectacle: namely, that of a socially awkward – albeit spectatularly wealthy – geek (like the one portrayed by Jesse Eisenberg in David Fincher’s The Social Network) get absolutely tarred and feathered. Twitter filled with jokes as the crowd grew impatient. Some of them were even good.

They underestimated Zuckerberg. Expectations for his performance before a series of committees of both houses of the US congress started out lower than subterranean. Yet even at the start, the 33-year-old billionaire did look absolutely terrified. Blinking vacantly in the strobe-flashes of the cameras, his expression while he sat listening to the senators’ seemingly-endless introductory remarks was not so much lost as “404 not found”.

But over the course of an often-agonising ten total hours of testimony before a joint sitting of the Senate commerce, science, and transportation committee, and the judiciary committee on Tuesday, and the House energy and commerce committee on Wednesday, Zuckerberg managed to come out not just unscathed but victorious.

In recent years, the Facebook CEO has made an effort to learn to be a more disciplined public speaker and a more responsive interviewee. On top of that, in preparation for this appearance Zuckerberg hired a crack team of outside consultants and lawyers to coach him, and even held mock hearings to hone his answers and manner, the New York Times reported. His investment paid dividends: Zuckerberg spoke with a glossy confidence and gave an effective and assured – though somewhat robotic – performance which left many of the lawmakers visibly charmed. He largely avoided answering questions he didn’t want to, and no lawmaker was able to press him to the point where he became visibly physically uncomfortable, as he has in the past.

It was possible to watch the Zuckerberg charm offensive play out in real time, not just on social media but on the financial markets. As soon as he began to talk, Facebook stock began to rise, and apart from a bit of a dip on Wednesday morning it pretty much never stopped. On Tuesday Zuckerberg’s confidence before the Senate committee gave Facebook shares their best single day of trading in two years, closing 4.5 per cent up. By the time Zuckerberg finished answering questions on Wednesday afternoon the stock price increase meant his own personal net worth had gone up by just under $4bn.

Far from the meltdown that many tuned in expecting to see, viewers were treated to Zuckerberg dealing patiently and even-temperedly with questions that occasionally betrayed a lack of even a basic conception of how the internet works, let alone Facebook. Some of his interrogators, especially in the Senate hearing on Tuesday, barely seemed to understand their own prepared questions even as they read them aloud.

This allowed Zuckerberg to get off considerably more lightly than he appears to have been expecting. A tantalising glimpse into the hearing we could have had was given to us when Zuckerberg accidentally left his sheet of notes open on the table when he left the hearing-room for a break. The notes, which were photographed, show that he was prepared for broader existential questions on subjects like workplace diversity and European privacy regulation which sadly, in the end, went largely unasked.

Instead, some lawmakers used their time to throw dozens of redundant questions to which we already knew the answers. Zuckerberg at times looked like he was struggling to suppress his obvious delight at answering questions which contained fundamental errors, causing howls of frustration on Twitter from the watching tech press, who understood the opportunity missed. Other times, lawmakers threw softballs, leading to such scintillating exchanges as the following, between Zuckerberg and Dan Sullivan, a Republican senator from Alaska:

SULLIVAN: Mr Zuckerberg, quite a story, right? Dorm room to the global behemoth that you guys are. Only in America, would you agree with that?

ZUCKERBERG: Senator, mostly in America.

SULLIVAN: You couldn't – you couldn't do this in China, right? Or, what you did in 10 years.

ZUCKERBERG: Well – well, Senator, there are – there are some very strong Chinese Internet companies.

SULLIVAN: Right, but you're supposed to answer “yes” to this question.

The main problem was the format didn't lend itself to a genuine search for insight. That's because any time it got half-way interesting, such as in an early exchange with South Dakota senator John Thune on the technical and linguistic difficulties involved in teaching AI bots how to accurately spot hate-speech, the dialogue would be abruptly cut off as each successive legislator ran up against their four-minute time limit.

Some legislators didn’t even bother trying to ask key questions about privacy and data protection, but instead decided to fawn or grandstand. Ted Cruz took an audaciously pompous line of questioning about how he felt Facebook was biased against the political right – without mentioning, of course that he actually ranked among Cambridge Analytica’s political clients.

The lack of coordination and preparation among his interlocutors allowed Zuckerberg time and again to cast Facebook as a company exists only to make people's lives better now and forever, rather than as a for-profit surveillance organisation. Time was wasted explaining over and over that, no, Facebook does not literally “sell data”, though John Cornyn, a senator from Texas, did pull off probably Tuesday night’s only true zinger with his muttered riposte: “well, you clearly rent it”.

There were some exceptions. California Democratic senator Kamala Harris, a former prosecutor, almost drew blood with a searing, sustained enquiry into whether there had been, when the company learned that user data had been shared with Cambridge Analytica, “a discussion that resulted in a decision not to inform your users”. In one of the few moments of the entire proceeding in which Zuckerberg found himself on the back foot, Harris pressed home the question a brutal seven times before her allotted four minutes were up.

His appearance before the House committee on Wednesday was testier in general but not much more enlightening. Anna Eshoo, a Democratic representative from California, scolded Zuckerberg for the opacity of the site’s terms and conditions, telling him: “you have to make it transparent, clear, in pedestrian language, just once, ‘This is what we will do with your data. Do you want this to happen, or not?’” Others pressed Zuckerberg for action controlling the sale of opioids on the Facebook platform. Zuckerberg nodded, smiled, and made the correct engaging noises at the appropriate times.

Despite his polish, the moments when Zuckerberg came closest to slipping up his mistakes were largely own goals rather than the result of incisive questioning. One particularly embarassing slip-up came during the Senate hearing when he accidentally answered “yes” to the question of whether the special counsel’s investigation into Russian interference in the 2016 election had served Facebook with subpoenas. Scrambling, he hastily muddied the waters a few moments later with: “actually, let me clarify that. I actually am not aware of a subpoena. I believe that there may be, but I know we're working with them.”

Mostly, though, Zuckerberg was poised enough to avoid any question he didn’t want to answer either by promising to “have people look into it and get back to you” or with a robotically careful line like “I am not specifically aware of that.” If faced with a tough question, he could simply run down the clock for four minutes until the questioner's time ran out. And the more he talked, the more Facebook stock soared.

In the end, the most interesting part of the hearing wasn’t what was said in the room itself but in watching it all play out on social media, where commentators from the two different worlds of technology and politics collided at the same real-time event. The conversation was split right down the middle into two distinct groups: those mainly frustrated and confused by Zuckerberg’s jargon-laden technobabble, and those mainly frustrated and confused by the lawmakers’ inability to understand the basic working principles of Facebook or even the internet – though mostly they agreed with each other on their distaste for Ted Cruz.

If nothing else, it was illuminating to see just how wide the gulf between those two worlds was.

Nicky Woolf is a freelance writer based in the US who has formerly worked for the Guardian and the New Statesman. He tweets @NickyWoolf.