The Problem of False Memory

Why reliance on eyewitnesses alone won't win your case

I had a hearing the other day involving a bus operator who was in an accident with a cyclist. "Preventable accident" cases are among the most common cases in transit industries. At the hearing, we all watched the video together. These buses are equipped with nine cameras; it's not impossible to miss something, but it's hard. In this case, the operator was in a dedicated bus lane. The video clearly showed the cyclist riding on the sidewalk to the right of the driver. There were two lanes between the bus lane and the sidewalk. The cars in those lanes stopped when the cyclist entered the crosswalk but for some reason, the bus driver did not. He didn't even slow down. He hit the front tire of the bicycle, causing the rider to fall off. Fortunately, he wasn't injured too badly.

Watching the video, the driver insisted that the cyclist had not been on the sidewalk prior to entering the crosswalk. In fact, he insisted that the bike had been traveling along his right side and then had traveled ahead of him and was gone by the time of the accident. He insisted on this, claiming that the video was a trick.

Not surprisingly, perhaps, I believed that the video was real and that the damage the cyclist and bike received were real. I found that there was a preventable accident and upheld the discipline. Notably, however, I did not think the bus driver was a liar. In my experience, people remember events very differently than what is captured on video or what is observed by other witnesses.

My experience is borne out by science. Even people with super recall tripped up by false memories, study finds. In one study, people who watched a staged robbery remembered it more accurately than the people who were the victim of the staged robbery. Another example: witnesses to a stabbing that took place in Times Square recalled that the assailant had a gun (he didn't) or that he'd acted in self-defense by wrestling a gun away from the police officer he stabbed (he didn't do that either). There were many witnesses and many videos. All the videos were consistent; the eyewitnesses were not.

How can this inform your arbitration practice?

• By being aware that it is well-established that eyewitnesses have false or incomplete memories of events, especially dramatic events, and figure out how to bolster the more accurate version of events without assuming that the witness is a liar. Asking for more specific details, like the time of day, other things they saw, heard, or smelled at the time, or whether they related their experience to anyone soon after it occurred.
• Review or request available video as soon as possible. Make sure the union is given the opportunity to review the video early in the case as well. A surprising number of employers either don't look at available video (and thus fail to conduct a full investigation) or don't share it with the union, who are then unable to assess the merits of their own grievance.
• Don't trust the video alone. While it may be more reliable than an eyewitness, it also needs to consider as a part of the total circumstances. In one case I had, the two employees who were alleged to have fought were so distant from the camera that the video the employer thought was a magic bullet was in fact useless. I had to figure out what happened from eyewitnesses after all but the video was still helpful for understanding the layout of the shop floor.
• When attacking a witness's credibility, go easy on the accusations of dishonesty if it's more likely the person is just misremembering. Poke holes in their story but suggesting that they are misremembering, rather than accusing them of lying. The arbitrator is looking for the better evidence, not the best evidence of all time. Ultimately, the story that makes more sense is the one that usually prevails.

A few other items

I know I must seem like a broken record about why I think AI should be treated very skeptically. Let's just say I was unsurprised to read that AI employees disfavor AI. Meet the AI workers who tell their friends and family to stay away from AI.

And as I predicted, unions are starting to negotiate language and win arbitration cases about AI usage. Unionized journalists at Politico just won a case concerning AI. Read about it here.

Finally, I just wanted to let folks know that I've decided to curtail my ad hoc arbitration practice to concentrate on writing and teaching. While I have some fun projects in the works, I also hope to update and expand my arbitration books and offer classes in arbitration practice. If you or your organization is interested in learning more about trainings I might offer, drop me a line: andrealdooley at gmail dot com.