Sensors and Journalism

Introduction & Executive Summary

This report you have opened is not monolithic. You can match its sections to your needs. The first section introduces our topic. It starts by describing the landscape where sensors and journalism combine, and continues on to define necessary terms for understanding this area of research. Reporters are using sensors in an era when the rapid development of technology is moving data into the mainstream of journalism. The increasing ubiquity of sensors, their increasing capability and accessibility are on the supply side, while investigative reporters, computer aided reporters and journalist/technologists are on the demand side. We are including drones in the field of sensing, partially because of the amount of attention they’re currently receiving, and partially because of their potential to extend human sight far beyond our bodily bounds. While recent commentaries about journalistic sensing have focused just on sensors that journalists have built themselves (or commissioned), our definition also includes journalistic uses of data from sensor systems that are not controlled by the reporters themselves. We have excluded opinion polling, information gathered by humans’ five senses, and data produced by monitoring computer processes like bit-torrent networks. That said, our description should not be used to separate sensor-based journalism from other reporting processes. The intellectual tools we discuss may be useful for many data-intensive projects, and sensor reporting needs to be integrated with traditional forms.

Then, scholar Charles Berret has written a chapter on sensor history, charting humanity’s efforts to extend the reach of our five natural senses. It starts with the scales unearthed by archeologists, the Neolithic markers like Stonehenge, and the agricultural tools from the Nile region. Berret notes that, in the 1500’s, the astronomer Tycho Brahe built a data network using the post, which compiled sensor readings to draw the most accurate and comprehensive star maps of his time. Cameras and sound sensors came in the nineteenth century, a moment ’when mechanical sensors were first treated with greater credibility than the human observer.’ The history outlined here only goes as far as the first half of the twentieth century, but within our time period it covers early medical sensors in the form of René Laennec’s stethoscope and Willem Einthoven’s electrocardiogram, and meteorologists’ use of doppler radar. The introduction finishes by outlining the characteristics of sensors that make them useful—or not—and helping readers identify what elements of the world can be sensed. The second section, containing case studies, examines seven projects that used sensors for journalism. Each study includes the story of what happened and then offers analysis in which we identify its distinctive or noteworthy elements, as well as the lessons journalists may take from the projects. The case studies start to show distinct types of sensor uses that suit different journalistic goals. The first type is when investigative reporters (environmental reporters in these two examples) design a sensing process to collect data with the intent of testing a hypothesis. They used relatively mature professional equipment and consulted with experts. They had justifiable confidence in their data, even though their processes were quite different from how scientists work when intending to publish in a peer-reviewed journal, or when doing work on behalf of regulators.

Another type of sensor use by journalists is accessing data from municipal sensor systems. The Sun Sentinel won a Pulitzer Prize by using tollgate data to investigate widespread speeding by off duty Florida police. The Washington Post published an extensive explanatory feature based on data from a network of microphones installed by the city law enforcement. A separate type of journalistic sensing involves DIY hardware development. At the moment, these projects value participation and informal science education. At the moment the equipment they use is unlikely to produce data that can be heavily relied upon in legal or health settings. However, the makers in this part of the field see great long-term potential, inspired by open-source software, a phenomenon that has returned great value to newsrooms. In case studies, we have also added analysis of the U.S. drone journalism industry, as it stands right now. At the high end, a small number of organizations are using footage shot by specialist pilots of professional cinematography drones. In the middle, enterprising news industry employees are experimenting with pro-am equipment costing hundreds, not thousands of dollars. Mainstream media organizations are also sourcing drone footage shot by hobbyists. All of this activity is proceeding despite a rapidly changing, highly contested regulatory environment. For the next section, about laws and ethics for reporting with sensors, we recruited 12 experts to write a chapter each. They applied their considerable knowledge and ability from professions in law, technology, ethics, academic research, and the sciences. In the individual essays, each helps identify and navigate the key issues that arise when their field intersects with sensors and journalism. The authors who address the privacy and surveillance issues write that these are early days for the field. The courts have thus far dealt with consent to record, defamation and false light in the context of cameras and microphones. The potential for journalists to break those laws using different types of sensors certainly exists, and if legal claims are made courts will

likely consider the ethical standards that emerge in these next few years of sensor reporting. The field is emerging even as relationship between newsrooms and their audiences transforms. Our authors suggest that journalists should involve their communities as they negotiate the tricky questions of who owns and controls personal data from sensors. Newsroom managers who have staff making or acquiring hardware should also acquaint themselves with the basics of open-source licensing. Often, journalists who design their own sensing systems will lean towards sharing their work under open-source principles, but this may involve legal liability if hardware goes wrong and causes physical damage. The risks are avoidable, however, as the article by Diana Cooper makes clear. Still in the realm of legal issues for hardware makers, this current phase of rapid, widespread DIY development is moving a lot faster than The Federal Communications Commission. The FCC requires that any electronic device that might emit radio interference be tested and approved before marketing. However, that regime did not consider many conceivable journalistic uses of custom sensors produced in small batches. If and when sensing, in particular drone use, becomes a widespread journalistic practice, human error is likely. Serious mistakes will attract negligence claims, following in a tradition codified by The Digest of Justinian in the mid-sixth century. It contained a section on ’Those Who Pour or Throw Things Out of Buildings’. The laws extended to falling things, as well. Despite this history, the novelty of drone journalism will make insurance tricky and expensive until the industry has more data on which to model risk profiles. The last group within the legal and ethical section concerns truth and accuracy. Sensors may seduce journalists into thinking their output is objective and free from the errors inherent in human testimony. That is a risky belief. We have drawn on the expertise of the EPA to show how reporters might design a sensor-based data collection process to improve their accuracy. For journalists, the concept of ground-truthing—supplementing sensor

information with human input—will be valuable. It should help introduce nuance, guard against mistakes and treat fairly the people at the heart of our stories. For the final section, we have distilled this report into a set of recommendations, including groups of strategic moves, good work practices, and efforts the industry may collectively consider. Strategic Recommendations

  • Identify and cultivate sensor sources for the beats you’ve prioritized.

  • Put a watching brief on open source sensing systems.

  • News nerds should do hardware too. Work Practice Recommendations

  • Before sensing, articulate your hypothesis.

  • Work with experts on complex stories.

  • Understand the entire pipeline for your story’s sensor data.

  • Combine sensing with traditional reporting. Recommendations for the industry, collectively

  • Journalists have an opportunity and a responsibility to report on sensor systems.

  • Advocate for access to data from publicly funded sensor systems. However, even though we have documented significant amounts of journalistic sensing here, we hope that this report will need updates as newsrooms keep combining their reporters’ ideas with new sensing opportunities.