Newsletter

The Sixth Conference on Computers, Freedom, and Privacy


Notes on Falsifying Electronic (Audio/Visual) Legal Evidence

By Ben Gross

Jordan Gruber and Tito Poza provided background on legal aspects of falsifying electronic evidence as well as information about related forensic tools.

Jordan Gruber said he got started in the field by wondering about the question if we will soon get to the point where we can falsify and audio of visual material so it can be completely falsely generated. Tito Poza talked mainly about tools the technical expert can use to evaluate the evidence.

The main themes are that there has been growth, pervasiveness, and significance of audio-visual recordings in the courts and society in general. The standards by which the recordings have been judged have actually weakened over the last forty years. Previous editing and falsification ability skyrockets due to digital technology. Forensic experts called on for assistance are routinely unaware of, or unwilling to acknowledge, their limitations in the area. This leads up to the fact that there is no easy solution.

The real problem is that we may be faced with a situation where the routine acceptance of electronic AV evidence leads to the inadvertent acceptance of competently falsified electronic AV evidence in an unacceptably large number of cases.

This is hard to fix for a number of reasons. First, electronic evidence usually brings us closer to facts. Our only tools are awareness, changing standards, and burdens of proof in the courts. To the degree we tighten standards, an unacceptibly large amount of perfectly good electronic evidence may be kept out. But simply trusting finder-of-fact may not be good enough where evidence is so pervasive. The solution is more awareness and starting thinking about standards. It has not yet happened where any large settlement or death penalty in a case where the evidence was later found to falsified.

The American legal system has moved in the last 50 years to the point where any evidence is let in and the jury can decide its value. A judge could theoretically not let the recording in to court. To get rid of the evidence you have to prove not that the evidence might have been falsified, but that it was falsified. It is much easier to answer a challenge to evidence than it is to scientifically prove the evidence as authentic. Tape authentication came to the forefront with Watergate tapes. It triggered the defendants to the possibility of falsified evidence. Tape authentication should really be called tape verification. The system is very much favored in terms of the proponent as oponent can not show that the evidence may have been or could have been falsified but that it has been falsified. Over time there has been a movement of trusting juries with evidence.

Legal video is creeping into every aspect of the courtroom. We live in such a media dominated society and world that audio and video are given regard above many other kinds of evidence. In the video world there is the silent witness theory. The recording is said to "Speak for itself." This is a long way from the seven pronged standard. The judicial system is getting more and more used have less and less burden of proof.

The forensic expert's proper role is technical authentication not legal authentication. There are technical and scientific limitations on experts. For example, experts admit they can fool other experts, so that means that expert could in turn be fooled.

Having the original recording equipment gives the examiner far greater ability in examinations. There is not a lot of legal precedence in forcing the original equipment to be used. Lack of access to original equipment has not caused the evidence to be discounted. The best weapon that challenger is to pin down the protocol that was used to create it.

With sufficient resources it is possible to look for second order effects which may be considered to be indicative of a problems ie. multiple 60 hertz tones. The amount of money that went into making the tape and the amount of money that went into examining the tape have a relation. Astronomical amounts of money went into examining the Watergate tapes.

Signatures may even vary on the same equipment depending on the state of the tape, how the equipment was used. ie. how the tape was stopped. Originality of record stop signatures is often called into question. The erase head is wider than the record head. This can come to light with magnetic development where the tape is examined for its magnetic properties, however it is easily circumvented and can be reproduced. In most cases only audible anomalies make it passed the initial stages.

Voice prints are a technique which give a digital "print" of a persons's voice. A lot of claims were made about the technique of voice prints in its early stages and then shot down and then let in. Now it is up for grabs. The bottom line is that under ideal conditions this technique works fine, but so does listening. Under forensic conditions it can be highly suspect. The FBI will not testify based on the evidence of voice prints.

Tito said he is constantly impressed with the processing ability of the human brain and ear. Tape enhancement in not usually a big deal. Certain kinds of noise are easily removable. Even in cases where you can not get rid of the noise you can often reduce it. There is a legal question of whether or not enhancement can change the speech elements. Additionally, There is a religious argument on whether or not you loose anything when you go to digital.

The speakers described what you need to competently falsify audio evidence and gave a few rough examples of forged speech done on home computer equipment. First you need, the right equipment (ie. a computer with spectrographic and waveform capabilities.) You also need knowledge of speech and forensics as well as enough of the target's speech previously recorded to do something with it. Their is a component of patience, artistry, and luck. It is always much harder to add material than to take out.

In summary: Don't cry digital wolf. Do not become too enamoured of digital technology as it is still generally hard to produce competent falsifications. It is easy to make mistakes that a forensic expert can find and the more complex the attempted falsification, the harder it is to pull off. There is a danger that bias inherent in a subjective process will come out if there are not enough comparisons of samples.


[ CFP96 Newsletter | CFP96 | CFP | general info ]

Comments and bug reports to Daniel C. Stevenson