When Vince sat down with the scientists from Xomnia, they suggested that because the team was working on such an old case with missing data, the puzzle of the August 4, 1944, arrest would almost certainly never be complete. Yet at some point the program’s algorithms should be able to predict what or who was the likely suspect.
To organize the massive amounts of data collected from documents and interviews, Vince developed a number of investigative initiatives. He called them the Residents Project, the Statements Project, the Media Project, the Mapping Project, and the Arrest Tracking Project.* The initiatives required hundreds of human work-hours, done mostly by a group of dedicated researchers, many of whom were volunteers and students. They ranged from teenagers, such as a student from Italy, who translated Italian press articles, to a retired Dutch professional in her seventies.
In addition to documents and book scans, the speech recognition portion of the Microsoft AI program was able to convert video and audio recordings to text, make them searchable, and translate them into English. As the team had hoped, the program began to show connections among people, addresses, and dates. These connections—policemen on the same raids, female informants who had worked together—had obviously been there all along but had not been noticed. Now the links began to form a narrative.
Because it was web-based, the AI program could be used anywhere. Pieter described the thrill of working with it at the National Archives: “If for instance, an address of interest came up in one of the files I was examining, I could very quickly cross-reference it within the database. Running the address through the AI would provide me with all relevant documents or other sources in the data store in which this address was mentioned. Sources where it was mentioned the most would appear highest. It could also give me a graphic on how this address connected to other relevant items such as different people who were somehow connected to this address. It could provide a map with all connections between this address and others and would indicate which connections were the most common. It would also provide a timeline when and where this address was most relevant.”
Investigative psychologist Bram van der Meer was approached, and he agreed to work with the team. Vince knew him as a criminal profiler and investigative psychologist in the Netherlands, who advised investigative teams throughout Europe and had worked on several cold cases. The team eventually brought him all the data they’d collected about the witnesses, victims, and persons of interest and asked him to examine it from a behavioral perspective. This included information about their backgrounds, family life, social and work life, and especially behavioral responses and decision-making in unusual situations or under specific circumstances.
In the hope of somehow making a miraculous discovery of physical evidence, the team developed a plan for evidence analysis with Detective Carina van Leeuwen, a cold case forensic detective with the National Police Corps. Since the investigation was not officially sanctioned, Vince knew that access to governmental laboratories for testing of physical evidence （e.g., DNA, fingerprint analysis, radiocarbon dating） could be difficult, but he was optimistic. As he put it, in the Netherlands there is probably no unsolved crime closer to the national heart than the betrayal of Anne Frank.
Another of the team’s investigative tools was straight out of a millennial’s playbook: crowdsourcing. From the day they announced the project and appealed to the public for any information they might possess regarding what had caused the Annex raid, the team received a steady flow of tips. Some even led to new theories that needed to be investigated; others were from people claiming to be the reincarnations of resistance workers or insisting that Anne Frank had survived the war and was living with a new identity somewhere in the world.
The investigation was deeply serious, but there were humorous moments. For instance, at one point Vince was amused to find himself taught a valuable lesson by a teenage student doing a school-sponsored day internship at the Cold Case Team office. He’d asked the student to confirm the addresses and telephone numbers of particular witnesses by checking these in a 1963 telephone book. He went over the names and addresses and explained what he was looking for. Then he asked the young man to repeat the instructions, which he did. “Any questions?” Vince asked. “Just one,” said the student. “What’s a telephone book?” The lesson: “Never presume.”
Based on existing theories, those newly developed by the Cold Case Team, and those received from the public, the team ended up with roughly thirty different possibilities of why the raid occurred. Several of the theories had already been heavily researched, but cold case protocol required making a due diligence review of the material, checking the source of information for accuracy, and carefully evaluating conclusions.