Homicide in the age of Amazon: How smart speakers and wearables are revolutionising detective work
by Thomas McMullan
In April 2017, Richard Dabate was charged with the murder of his wife, Connie. His story – that a man who looked and sounded like Vin Diesel had broken into their home, attacked him and shot his wife – was undermined after detectives pulled data from the victim’s Fitbit. Richard Dabate’s version of events, built on panic-alarm notifications and work emails, was disassembled by a very different timeline, based on the amount of steps Connie Dabate took towards her death.
Earlier this year, in a separate case, data from an Amazon Echo was used as part of a murder investigation in Arkansas. After Victor Collins’ body was found in his friend James Bates’ pool, prosecutors pointed to information from the accused’s Nest water meter to argue that someone had used the garden hose to clear water off a patio. The authorities’ attempts to further access information from Bates’ Amazon Echo – which may have picked up recordings of the crime via its built-in microphone – were blocked by the company until Bates himself gave permission for the evidence to be handed over.
The hope, at least for Bates, was that the Amazon Echo could be an alibi; that data from the smart speaker would exonerate him of the murder of his friend. The court case is ongoing.
There’s an increasing number of sensors on our wrists, on the walls of our homes, in our pockets, on our mantlepieces, and these machines tell stories that detectives can use to build, or break apart, investigations. The rise in the Internet of Things is leading to a new era of police work, where the established fields of forensics and witness testimony are coming up against prevalent data gathering, voice recognition and movement monitoring.
“Increasingly, with the Internet of Things, there will be data streams which could be analysed forensically, and which could piece together an investigation,” says Professor Allan J Brimicombe, a director of the British Society of Criminology and head of the Centre for Geo-Information Studies at the University of East London.
“Say, for example, a burglar breaks a window to get into a house. The temperature in the house drops. Well, that’s going to be recorded, so you would be able to fix the time of burglary much more accurately than with someone who says ‘I left the house at 10am and came back 10pm to find the window broken and things gone’.”
This ability for sensors to pinpoint times and locations could be the Internet of Things’ greatest contribution to investigations. Brimicombe adds that if you ask ten people what happened in a car accident, you’ll get ten different stories. “Whereas the data is the data. If cars had black boxes, you may get a more objective take on what happened.”
Reading evidence from the data
Autonomous cars could change things with onboard computers capable of logging a vehicle’s every lurch. Yet if the proliferation of in-built tracking devices really does become the new backbone of police work, what does that mean for the way investigators handle testimony? In a city of a million sensors, what happens to human witnesses?
In a city of a million sensors, what happens to human witnesses?
Marion Oswald, senior fellow in law at the University of Winchester, believes detectives won’t be directing all their questions to smart homes just yet. “The evidence is not the pure data from the technology,” she stresses. “It’s evidence given based on the interrogation and interpretation of the data log, so witness testimony will remain important. In fact, it might become even more important as the way these technologies generate data is complicated and opaque.
“For instance, we’ve recently seen a case in Ohio where data from a pacemaker was used to prosecute a man for arson,” she notes. “But this case needed a cardiologist to interpret the heart-rate data and compare it to the defendant’s explanation of what he was doing at the time.”
In that example, a cardiologist might have been able to build a narrative using medical information, but the increasing frequency of home hubs, fitness trackers and smart meters in investigations requires a technical skillset that might be further out of reach for many detectives and lawyers. Brimicombe suggests that a “new breed of IT expert witnesses” is possible, but there are likely to be multiple barriers to any standardisation of this in court. As Oswald points out: “You will have to provide robust explanations of how the system works, and these technologies are changing and developing all the time.”
There is also the issue that the Internet of Things is far from tamper-proof. Alice Hutchings, a senior research associate at the Cambridge Computer Lab, tells me one important issue in this area is the vulnerability of tech-heavy systems.
“A challenge for prosecutors is likely to be proving that evidence from such devices has not been interfered with”
“Mirai botnets use insecure ‘security’ cameras and similar devices that have been set up using default usernames and passwords,” she explains. “This means devices can be harnessed by criminals for denial-of-service attacks and other unauthorised activities. A challenge for prosecutors is likely to be proving that evidence from such devices has not been interfered with. Another example comes from the recent WannaCry ransomware, which reportedly affected speed cameras.”
The case Hutchings mentions happened in June, when around 55 traffic cameras in Victoria, Australia were infected with the same malware that had previously caused havoc for the NHS in the UK. This example wasn’t the result of a targeted attack and instead happened because of a contractor that mistakenly connected infected hardware to the cameras. All the same, it shows the potential for sensor systems to be undermined by cyber-attacks and their data stolen or manipulated by hackers.
“Part of an investigator’s toolkit”
Regardless of its flaws, Hutchings thinks evidence from devices such as home hubs and fitness trackers “will become a standard, and quite an important, part of an investigator’s toolkit”. Marion Oswald too believes that, even though its legitimacy in court isn’t assured, evidence from these technologies will become more common.
“Indeed, I might even say that it would be negligent of law-enforcement organisations not to consider the potential of these technologies,” she tells me, adding: “They need to consider if these technologies are a help or a hindrance.”
For instance, if data from a Fitbit is so open to interpretation that it doesn’t prove anything, focusing on such data might be a distraction from other forms of evidence.
“On the other hand, there might be situations where data within IoT was crucial to help the police build a case,” she adds.
Much of this is breaking new ground, and there will be – as Brimicombe says – “an inordinate amount of precedents people are going to have to think about, and deal with through the court”. What’s needed in the meantime is an open discussion about what is and isn’t an appropriate use of potentially sensitive information. You can’t get much more intimate than a pacemaker, and while data from that device may be allowable in an arson case, what about a minor offence, such as theft? What about divorce proceedings, to prove adultery? And should the same rules apply for a heart-rate monitor on an Apple Watch?
Rules and rights
Apple is a looming figure here, namely in terms of its 2016 conflict with the FBI, which saw the tech giant being fiercely criticised for not granting access to the iPhone of San Bernardino shooter Syed Rizwan Farook. Apple said no, with the FBI eventually bypassing the company’s security system, reportedly with the help of Israeli company Cellebrite.
In contrast, Amazon did hand over information to aid the investigation into Victor Collins’ murder, but only after the Amazon Echo’s owner – defendant James Bates – gave the go-ahead. In an age of Snoopers’ Charters and encryption wars, the rules and rights around who actually owns and controls the information recorded and stored by Internet of Things devices is still very much in flux. When I reached out to Fitbit to comment on the Connie Dabate case, a spokesperson said the company wouldn’t comment on a particular story, but that “like many other companies, Fitbit may comply with valid law-enforcement requests”. The spokesperson added that “respect for the privacy of our users’ data drives our approach to complying with legal process. When we receive a request, our team reviews it to make sure it satisfies legal requirements and Fitbit’s policies.” Amazon did not respond to a request for comment on the Victor Collins case.
One thing is certain – the number of sensors in our lives is increasing
Despite the legal and ethical arguments about which data is fair game for investigators, one thing is certain – the number of sensors in our lives is increasing. The worldwide market for smart speakers such as the Amazon Echo and Google Home is expected to grow to a staggering $13 billion by 2024, while the market for wearables such as the Fitbit and Apple Watch are growing beneath a shifting focus on health and fitness.
With a microphone in almost every home, and a step counter on almost every wrist, there’s going to be a lot of data collected, and a lot of potential evidence for detectives.