Despite great variation among different countries’ legal systems, some challenges — such as problems with forensic science and eyewitness testimony — stretch across borders.   

The U.S. National Academy of Sciences and the United Kingdom’s Royal Society recently held a meeting in London to explore ways to strengthen the role and reliability of the science used in the two nations’ court systems.  

Among the speakers was Niamh Nic Daéid, a forensic science professor at the University of Dundee in Scotland, who offered an overview of forensic science’s past and future.  

Nic Daéid noted that some TV programs and novels feed an unrealistic vision of the field’s capabilities. “It builds the expert into an omnipotent font of knowledge about pretty much everything, and that we will provide the solution,” she said.  

It’s an expectation of certainty and flawlessness that forensic science can’t fulfill. Forensic scientists have come under pressure to do things faster, better, and cheaper, said Nic Daéid — a combination that sometimes doesn’t generate the robustness that courts require. Moreover, the scientific foundations of some forms of evidence have been increasingly scrutinized.  

The 2009 National Academies report was “a specific fundamental moment within the forensic science community where some of our challenges were laid bare,” continued Nic Daéid. “Some of what we deliver was deemed to be not valid and reliable, in terms of the scientific robustness one would want to have underpinning it.” 

Part of the solution is about accreditation, quality assurance, and setting standards — efforts England and Wales have made some progress on in recent decades, noted Nic Daéid. Nevertheless, she said, “the challenge that we have been faced with in 2009 has not gone away.” In 2016 and 2019, for example, further reports in the U.S. and U.K. voiced serious concerns about the scientific validity of some of the forensic science appearing in the courts.  

“We now appreciate the problem across all of our community, but we’re still a bit stuck,” said Nic Daéid. “And we’ve got opportunities of either continuing to be stuck, or to try to look at things from a different perspective.”  

Key to making progress will be facing four challenges, she argued — the challenge of coping with the pace of change, including the digital revolution; the challenge of strengthening and opening up research; the challenge of communicating clearly in the courtroom and to the police, including about limitations and uncertainties; and the workforce challenge, which will require breaking down silos both between the legal and scientific communities and among various parts of the forensic science community.  

“The opportunity for us, as a community together, to think differently and take ownership of our challenges is right within our grasp,” said Nic Daéid. 

Improving eyewitness testimony

Science has shed light on problems with another type of evidence commonly used in courts on both sides of the Atlantic — eyewitness testimony.  

“In the last 20 to 30 years, there have been several thousand cases in the United States in which judges have determined that previously convicted defendants were, in fact, factually innocent,” said Jed Rakoff, a senior judge of the United States District Court for the Southern District of New York, and co-chair of a 2014 National Academies study on eyewitness identification. While there are multiple reasons for this, “statistically by far the biggest element has been inaccurate eyewitness identifications,” said Rakoff.  

Ray Dolan, a professor of neuropsychiatry at University College London, explained how flaws in human memory can sometimes lead to faulty testimony. For example, if during the process of recalling an event, a person is provided with misleading information, this can contaminate their memory, he said. In fact, the mere act of imagining an event can lead to contamination between what the person was asked to imagine and what they actually experienced, as the same regions of the brain that are active in imagination are also involved in memory.  

It may be possible to develop a model that could help predict the likely accuracy of an eyewitness’s testimony, said neuroscientist Thomas Albright of the Salk Institute for Biological Studies, who co-chaired the National Academies’ study with Rakoff. With enough data on how certain variables — the cognitive state of the witness, for example, and the lighting at the crime scene — are related to accuracy, a model could make predictions about how likely a particular eyewitness’s testimony is to be correct.  

Navigating new forms of evidence

In addition to long-standing challenges, the meeting also explored new types of evidence that are making their way into courtrooms.  

Eoghan Casey of the University of Lausanne spoke about challenges in understanding and using digital evidence — for example, location data from cell phones — in the legal system. “The questions in courts are starting to come, as they should, as we don’t have the science to back up a lot of the reliability questions about digital evidence.”  

He noted a recent Massachusetts Daubert hearing to determine the admissibility of iPhone frequent location data in a homicide case — data that was claimed to show the defendant near the crime scene. The court decided that the evidence didn’t meet the Daubert standard for reliability and therefore could not be presented in court.  

The biggest problem is the lack of testing and limited experimentation, which is only done when the courts ask for it in a specific case, said Casey. “We’re not doing the scientific research proactively, we’re doing it reactively, and too late.” 

“Essentially, any location-based evidence that’s pulled out of mobile devices or telecommunications systems are being presented like fact, when in fact they’re quite uncertain,” said Casey. “In a scientific context, I think we can start to address the uncertainty and really start to use that information in appropriate ways.”  

Gillian Tully of King’s College London offered an introduction to likelihood ratios as a way to approach and express some of the uncertainties in forensic evidence.  

“There is uncertainty associated with every scientific measurement that you make, or every evaluation that you do, and the magnitude of the uncertainty depends on what it is you’re measuring,” Tully said. 

With cell phone location data, for example, there may be uncertainties that come from the direction of the cell towers and whether there were buildings and trees between the tower and the site of interest, and what cell phone network the suspect was on.  

“Probabilities enable us to reason logically when there is that sort of uncertainty, and probabilities aren’t binary,” said Tully.  

She also stressed that it’s important for forensic scientists to be transparent about the assumptions they are making and the extent and quality of the data on which they are basing their evaluations. 

“I do not think that we as forensic scientists are transparent enough,” she said. “So that’s really important for us to address, and that can partially be addressed through rules and standardization, but it also needs a lot of education.”  

The meeting is planned to be the first in a series of joint meetings about science and the law in coming years.