Latest / CLEs You Actually Want to Hear / Underlying Biases in Consumer Litigation: Identifying and Challenging Hidden Assumptions
Transcript
- 0:03Hello, everyone, and welcome to today's program,
- 0:07Underlying Biases in Consumer Litigation, Identifying
- 0:10and Challenging Hidden Assumptions. I'm Alapoli
- 0:13Castro, Director of Education and Professional
- 0:16Development at Beverly Hills Bar Association,
- 0:18and I am delighted to introduce today's speaker,
- 0:21Molly McKibbin. Molly is a partner at Green Berlet
- 0:25and Wheeler in Los Angeles, where her practice
- 0:27focuses on catastrophic personal injury and wrongful
- 0:30death cases. Her many accolades include being
- 0:34chosen as Consumer Attorneys of California's
- 0:37Consumer Attorney of the Year. She's on the board
- 0:40of CAOC and Consumer Attorneys Association of
- 0:43Los Angeles, and is a member of the American
- 0:46Association for Justice, the Association of Business
- 0:49Trial Lawyers, and the Women Lawyers Association
- 0:52of Los Angeles. Angeles and the Academy of Truck
- 0:54Accident Attorneys. She's been published in many
- 0:57publications including the Daily Journal, the
- 1:00Advocate and Forum Magazine, and volunteers for
- 1:02the Sojourn Domestic Violence Legal Clinic. We
- 1:05can't wait to hear what she has to share with
- 1:06us today so I'm going to turn it over to her
- 1:08now. Good morning everyone. I'm going to go ahead
- 1:12and share my screen and make sure everyone can
- 1:20see that. All right. For the display again. All
- 1:26right. Is that better? Can you try again? Okay.
- 1:39Let's see. Maybe if I end and start it again.
- 1:45I am here to talk to you about bias. Some of
- 1:48you are probably here because your MCLE reporting
- 1:50period is soon, but I hope that my presentation
- 1:54gives you something to think about, even if you
- 1:57don't necessarily do consumer litigation and
- 2:00examining the biases that we may have in litigating
- 2:04cases, representing clients, assumptions we make
- 2:07about. our client's claims, our client's defenses,
- 2:11or the evidence in the case. And so the goal
- 2:15here is just to get you thinking about bias in
- 2:18a different way. So bias is something that we
- 2:22as attorneys and trial attorneys should be intimately
- 2:25familiar with. It's something if you are in front
- 2:28of juries that you discuss with them, right?
- 2:31Your goal is to do a liberal and probing examination
- 2:35of all the jurors that are in your panel to decide
- 2:40if they are the right jurors for your case, if
- 2:42they have any biases or prejudices which might
- 2:45affect their ability to be impartial. And this
- 2:48This is such an important concept that some states
- 2:51have laws that actually preclude a inhibiting
- 2:58a lawyer in order to do that liberal and probing
- 3:02examination. And so every lawyer has their own
- 3:05example that they will use during vaudier to
- 3:07explain what the concept of bias is so that jurors
- 3:11can understand it and they can recognize, you
- 3:13know, hey, I've been in a car accident myself
- 3:15and I might not be the most impartial juror on
- 3:18a car accident case. And so There are biases
- 3:23that the people and groups who create information
- 3:26that we rely on form assumptions that we or the
- 3:29other side can make about parts of our cases.
- 3:32And we think about it as it relates to the jurors,
- 3:34but oftentimes we don't think about our own biases
- 3:37that we have throughout the course of our client's
- 3:40case. And we can't always trust our brain, right?
- 3:44Biases are simply things that are based on stereotypes.
- 3:48And stereotype is when our brain is making an
- 3:53assumption, making a judgment based on an assumption
- 3:57rather than on actual knowledge. So we perform
- 4:01unconscious classifications of people or things
- 4:04based on our life experiences. We can classify
- 4:07people by age, gender, race, or other stereotypes.
- 4:12And these cognitive shortcuts actually result
- 4:15in prejudgments that can cause us to make rash
- 4:18decisions. Biases can be based on inaccurate
- 4:24information. Sometimes bias can be a good thing,
- 4:26but oftentimes it can be negative. It can perpetuate
- 4:29systemic inequality. It can prevent our clients
- 4:32from receiving fair and equal treatment, whether
- 4:34they are a plaintiff or a defendant in a case.
- 4:37They can result in a client receiving less compensation
- 4:41than they are owed under their circumstances
- 4:44based on the harms that they have suffered. And
- 4:47it can cause us to overlook potential avenues
- 4:50of recovery for our clients. And so when left
- 4:53unchecked, these biases can take over, right?
- 4:57They can cause us to become blind to things that
- 5:00we should be aware of. So this is a... tweet
- 5:04that I thought was especially topical. Half of
- 5:09the United States and UK population are women.
- 5:13And the study that was performed was found that
- 5:16women are less likely to have CPR performed on
- 5:19them because men are afraid of being accused
- 5:21of being inappropriate. Women are also experienced
- 5:26bias in a variety of other ways. For instance,
- 5:28the Apple health kit that was available on the
- 5:31iPhone is still available on the iPhone. had
- 5:34specialized tracking that was made available,
- 5:36including what your copper intake was. But it
- 5:40wasn't until iOS 9, so nine versions of the operating
- 5:44system of this phone, that a period tracker was
- 5:46available, because there was an assumption that
- 5:48that was not something that a 50 % of the population
- 5:52would be interested in. Whoops, I'm sorry about
- 5:54that. it's important for us to examine these
- 6:01biases and the information that we're relying
- 6:03on or the life experience that we have, right?
- 6:06Many of the men that were in the study likely
- 6:09had some life experience that made them feel
- 6:11like they would not feel comfortable or thought
- 6:13it would be inappropriate to perform CPR on a
- 6:15woman that they didn't know. A study that was
- 6:19performed recently on Google search engine found
- 6:22that it creates an echo chamber for negative
- 6:26stereotypes. So for instance, when you type in
- 6:29the phrase are women into Google, the first things
- 6:32that pop up are, are women evil? Are women allowed
- 6:37in combat? Or lastly, are women attracted to
- 6:40money? Not are women smart? Are women capable?
- 6:44Are women in Congress? These are the things that
- 6:47Google is automatically populating for a person
- 6:50who is doing a search on the internet. And speech
- 6:54-to -text tools, there was a study that was done
- 6:56in 2020 and found that speech -to -text tools
- 7:01that many of us use in our day -to -day life
- 7:03were mistranscribing African -American voices
- 7:08at a rate almost double of those of white people.
- 7:12And so these are just, again, ways in which the
- 7:16world is created based on stereotypes, based
- 7:19on biases. And we have to think about how is
- 7:22that going to apply to our practice and our representation
- 7:25of clients. This tweet just made me laugh, but
- 7:29it is absolutely true. So for one important thing
- 7:32I think to understand, if you are in... litigation
- 7:35that involves medicine in any way, whether it's
- 7:38a medical malpractice case, whether it's a pharmaceutical
- 7:41case, whether it's a personal injury case or
- 7:43wrongful death case, you have to examine the
- 7:46information on which your experts, the treating
- 7:49physicians, are relying. And where is that coming
- 7:53from? So the public health system in the United
- 7:55States has been historically biased towards the
- 7:58male perspective. In 1977, the FDA actually recommended
- 8:02that women of childbearing years be excluded
- 8:05from all clinical research. And because of this
- 8:07gendered bias, many women now experience delayed
- 8:11diagnoses, misdiagnoses, and suffer from adverse
- 8:15drug effects. Because if we're not included in
- 8:17clinical trials for a medicine, it's very likely
- 8:20that that medicine and the makers of that medicine
- 8:22have no idea how it will affect women ages 16
- 8:26to 45. And so it's unsurprising that eight of
- 8:29the 10 of the drugs that were removed from U
- 8:34.S. markets in the 90s were removed almost exclusively
- 8:39because of adverse effects that were suffered
- 8:41by women. And it wasn't until 1989 that the National
- 8:45Institute of Health amended its policy to include
- 8:48women and minorities in research studies. But
- 8:51it wasn't until 1993 that this policy became
- 8:54federal law. And it wasn't until 2016 until the
- 8:58National Institute of Health actually required
- 9:01that sex be considered a variable in these clinical
- 9:06trials. So it is unsurprising then that In 2015,
- 9:11there were five times more studies on erectile
- 9:13dysfunction than there were on PMS, even though
- 9:16only 19 % of men's experience erectile dysfunction,
- 9:20and 95 % of women experience PMS. Cardiovascular
- 9:25disease is the number one cause of death for
- 9:27women, but only 5 % of National Institute of
- 9:30Health funding goes towards research around coronary
- 9:34artery disease for women. And out of 112 internal
- 9:38medicine programs reviewed in 2016, 25 % of them,
- 9:44a quarter of them, did not include any instruction
- 9:47on menopause, and 70 % did not include any instruction
- 9:51on fertility. Again, these are just more examples
- 9:58of the information that exists in the world,
- 10:00the structure that has been created, the assumptions
- 10:03that people who do studies, that people that
- 10:05put information out make, and that leads to biases
- 10:09that can leak into our cases and assumptions
- 10:12that we make about our cases. So one of the first
- 10:15areas I want to talk about how these biases can
- 10:17affect our cases in representing consumers is
- 10:20workplace safety. Women are becoming more and
- 10:25more involved in construction. Prior to 2016,
- 10:28women accounted for roughly nine out of every
- 10:31100 employees in the construction workforce.
- 10:34However, since then, the female employment share
- 10:36has climbed nearly every year. at over 11%. And
- 10:41they're more involved in trades. The number of
- 10:44apprenticeships held by women has more than doubled
- 10:46from 2014 to 2022. And the number of women in
- 10:52apprenticeships has reached its highest level
- 10:54ever, over 300 ,000 women. Despite that fact,
- 10:59the equipment that many of these women are being
- 11:01given is not designed for their bodies. When
- 11:04a person participates in an apprenticeship program,
- 11:07whether that's for construction or electrical
- 11:09work or anything at all involving an apprenticeship,
- 11:13the program typically provides them equipment
- 11:15that they're going to need in order to do their
- 11:18job. But that has not been broadened to include
- 11:23equipment that is specifically designed for the
- 11:25female body. And we know it's very obvious that
- 11:28safety clothing and equipment helps avoid accidents.
- 11:30Things like harnesses, hard hats, fall protection
- 11:33systems, goggles, gloves. But it's designed for
- 11:37the average male body. So goggles do not fit
- 11:39their faces. Flame resistant clothing is too
- 11:42big or too long. Safety vests and harnesses are
- 11:46given in a men's size small, which still may
- 11:48be too big for a lot of women. And it makes sense
- 11:52then. that improperly fitting safety equipment
- 11:55can either contribute to or cause injuries. So
- 11:59female firefighters experience a four times greater
- 12:03rate of injury than men in part because of their
- 12:06ill fitting personal protective equipment. And
- 12:09that's an added obstacle that a female faces
- 12:11in order to become a successful female firefighter,
- 12:15right? Is that she has to pass all the same tests
- 12:17and do all the same things that her male counterparts
- 12:19do, but then she also has to maybe walk and learn
- 12:22to walk in boots that are a size too big or deal
- 12:25with clothing that does not properly fit her
- 12:27body and impedes her ability to run or carry
- 12:31a ladder. So, you know, some examples that I
- 12:35saw when I was researching this talk were women
- 12:38being given welding boots in a size men's 9 .5.
- 12:42And I know that a lot of women do not have a
- 12:44shoe that big. Being given gloves that were so
- 12:48huge that they couldn't pick up tools. And so
- 12:51what is happening in practical from a practical
- 12:54standpoint is women are taking their own money
- 12:56and having to go buy equipment, but many of them
- 12:59do not have the money to do that. So the lack
- 13:02of appropriate PPE can also present employment
- 13:05issues from an employment perspective. So looking
- 13:08for the correct PPE can delay a person's hiring
- 13:11or complaining about the fact that I don't have
- 13:14the properly fitting equipment can lead to retaliation
- 13:18or firing. And in fact, there were two lawsuits,
- 13:21one in 2017, where a police officer... sued her
- 13:28department after they refused to give her protective
- 13:30gear that would fit her body as she was pregnant.
- 13:33They refused to give her a properly sized bulletproof
- 13:36vest so she didn't have one and they denied her
- 13:39request to carry some of her equipment in her
- 13:41pockets and her vest to lessen the strain on
- 13:44her abdomen as she's required to carry a 25 pound
- 13:47duty belt. And there was a 2015 lawsuit in Louisiana
- 13:51where a shipyard employer failed to provide appropriate
- 13:56PPE for a female equipment or for a female employee
- 13:59and she sued alleging gender discrimination.
- 14:05You know this obviously extends to the military.
- 14:08I know not a lot of our cases involve the military
- 14:11but I have had many clients who have military
- 14:14experience or are current military members. And
- 14:17lifting and carrying heavy loads for long periods
- 14:20of time are a major risk factor for musculoskeletal
- 14:23injuries. And researchers say that carrying more
- 14:25than 10 % of your body weight increases the chance
- 14:29of injury exponentially. And so based on the
- 14:32army's height and weight standards for soldiers,
- 14:35a woman between the ages of 21 and 27 who stands
- 14:38five feet seven tall can carry up to 100 or typically
- 14:43weighs about 161 pounds, which means she should
- 14:45be carrying about 16 pounds of equipment. But
- 14:48when these soldiers go on road marches they have
- 14:51a 35 pound rucksack and in the field their rucksack
- 14:54can weigh up to 50 pounds so instead of being
- 14:57given equipment that fits their bodies these
- 14:59women are forced to adapt to ill -fitting gear
- 15:02and they are required to wear the gear that the
- 15:06army issues during their basic training and once
- 15:09they are assigned to a unit they can purchase
- 15:11other equipment but that is again another hurdle
- 15:13that these women face. This also bleeds into
- 15:18medical professionals. So the American Medical
- 15:22Association estimates that Over a quarter of
- 15:25doctors in the United States are women now. And
- 15:28this is up dramatically from the 1970s, where
- 15:31it was only about 5 % to 7%. 50 % of all medical
- 15:35school applicants are women. And the total proportion
- 15:38of practicing female doctors is only going to
- 15:40grow dramatically in the coming years. But again,
- 15:43most instruments are designed for male hands.
- 15:46Women's hands are male doctors. Women's hands
- 15:49are smaller than men's. A man's thumb size can
- 15:52be as much as half to 1 .5 times bigger than
- 15:56a woman's. Women have about half the grip strength
- 16:00traditionally and squeezing strength as men.
- 16:05And the consequences of that are that their smaller
- 16:08fingers will float in finger holes for forceps
- 16:12or scissors or things like that, potentially
- 16:14causing discomfort, but more importantly, potentially
- 16:16causing instability and affecting their ability
- 16:19to perform surgeries or procedures. And then
- 16:22instruments that require them to do extra squeezing
- 16:25is going to cause repetitive damage to their
- 16:28ligaments and joints and may... cause them to
- 16:31exit the workforce earlier because they may have
- 16:34a disability. This is a study that came out fairly
- 16:40recently where the things that I'm talking about
- 16:43in terms of ill -fitting equipment was confirmed.
- 16:47And this study actually drew on 67 hours of operating
- 16:51observation where the researchers actually observed
- 16:54female surgeons and compared them with male surgeons.
- 16:58And these surgeons report having all the same
- 17:01issues that I'm talking about. Everything from
- 17:04their surgical spaces to their tools were designed
- 17:06with them. male body in mind and made their job
- 17:09much more difficult than their male counterparts.
- 17:13Another way that bias can rear its head in workplace
- 17:18safety is the actual way that the environment
- 17:21that someone is working is designed. So a great
- 17:24example are airline cockpits. They are designed
- 17:27to fit the measurements and proportions of a
- 17:30male body. So in order to operate the equipment,
- 17:32you have to have a certain amount of strength
- 17:34and be a certain height. And a study that was
- 17:38done found that 70 % of female pilots could not
- 17:40appropriately reach the pedals, could not appropriately
- 17:43reach the flight deck controls, levers, and points
- 17:46of visibility within the cockpit. Another area
- 17:49that you see this issue is in truck and tractor
- 17:51cabs for commercial drivers. Women can have a
- 17:55difficult time reaching controls and getting
- 17:57the seats adjusted while still maintaining contact
- 18:00with their pedals. So this is also not only disadvantages
- 18:04women, but it also disadvantages people of color
- 18:09who may be smaller in size than the average white
- 18:12male. Another way that we see bias affect workplace
- 18:19safety relates specifically to race. So worker
- 18:23fatalities have increased year over year since
- 18:26about 2015. White worker deaths only increased
- 18:31from, in the timeframe from 2015 to 2019, the
- 18:34white worker deaths only increased about 1 .7%.
- 18:37But Hispanic worker deaths increased 20%. Black
- 18:41worker deaths also increased. by 28%, and Asian
- 18:45-American worker deaths increased by 59 % during
- 18:49that same period. And the question is often,
- 18:53why? Why is there that exponential increase for
- 18:55people who are not white? And one of the basic
- 18:59answers that research has shown is that workers
- 19:02of color are traditionally assigned the most
- 19:05dangerous jobs. And there are often language
- 19:08barriers for these workers to understand how
- 19:10to properly operate machinery, to understand
- 19:14instructions and warnings, and not having them
- 19:16in their native languages puts them at a disadvantage.
- 19:20They also have difficulty reporting workplace
- 19:22safety issues if they have immigration status
- 19:25concerns. So they might push through and work
- 19:29through dangerous working conditions. So if someone
- 19:32who is in charge of safety knows a group of people
- 19:35are more likely to be injured or die, right?
- 19:37If you have workers that are not white, they
- 19:39have an obligation that they... do everything
- 19:42they can to make them safe, whether that's provide
- 19:44them instructions in their native language or
- 19:46make sure that they are being given opportunities
- 19:49to work in roles that are not the most dangerous
- 19:52or the most likely to cause them an injury. Another
- 19:58area where we see where we see these prejudgments
- 20:03and stereotypes is in insurance claims. If you
- 20:06have a client that has a bad faith claim or whether
- 20:10it's first party or whatnot, this is an area
- 20:13where bias is very, very prevalent. So this state
- 20:19farm lawsuit, it was from 2022 and it was born
- 20:26out of a study that was done by a civil rights
- 20:31organization and a law firm that predominantly
- 20:35handled civil rights issues. The study found
- 20:43that black homeowners were having issues getting
- 20:47insurers to pay their claims. So the study focused
- 20:51on a nine month period from 2021 with about 800
- 20:54participants. And State Farm is the largest homeowner
- 20:58insurer in the United States. The study found
- 21:02that white customers were able to get their claims
- 21:06approved and with fewer than three interactions
- 21:08with the insurance company employees. And they
- 21:11were one third more likely to have their claims
- 21:14paid in less than one month versus black customers
- 21:17were 20 % more likely to have to talk to a representative
- 21:20on at least three occasions and much more likely
- 21:24to have to do extra. paperwork. So pretty stark,
- 21:29the experience for white insureds versus black
- 21:33insureds. And another way that insurance companies,
- 21:38and I think this is becoming more and more prevalent
- 21:40every day, is their use of AI. So State Farm
- 21:46uses specialized technology, as I'm sure many
- 21:50companies do. And they use it to help process
- 21:52their claims. And they use a software from a
- 21:55AI firm called FRIS that flags claims for potential
- 22:00fraud. And the way that FRIS decides that a claim
- 22:04is potentially fraudulent is that it gives each
- 22:07policy holder a risk score by running the customer's
- 22:10information through its algorithm and its computer
- 22:13programs, which analyzes the language and their
- 22:16claim narrative, as well as the customer's profile.
- 22:20And they provide the score based on demographic
- 22:23data about the neighborhood, such as the degree
- 22:26of urbanization, there are words not mine, crime
- 22:29statistics and data harvested from social media.
- 22:34So both of these things I think are important
- 22:37to keep in mind where you have a bad faith claim.
- 22:40and you're dealing with arguments of medication
- 22:43of damages or, you know, where your client's
- 22:48claims about having difficulty getting claims
- 22:50paid are not being believed. I think it's important
- 22:52to really delve into what was their personal
- 22:55experience, how, you know, how... did their claim
- 22:58process go forward? And then also think about
- 23:01what is the data that State Farm is relying upon
- 23:04or the insurance company is relying upon to make
- 23:07their judgment call about whether to approve
- 23:09a claim or deny a claim. One way that we here
- 23:16in California have... addressed bias and the
- 23:20reality of the world that our clients live in
- 23:22and that their cases are presented is in the
- 23:25area of lost earnings and there has been legislation
- 23:28that has actually changed the course of how the
- 23:33calculation of lost earnings was being done.
- 23:37So California barred the use of race, gender,
- 23:42or ethnicity considerations in experts calculating
- 23:45loss earnings. So it was the first law of its
- 23:49kind in the United States that was passed. It's
- 23:52California Civil Code Section 3361. And you can
- 23:55see it here on the screen where it says that
- 23:59you cannot, an expert is not to consider race,
- 24:02gender, or ethnicity and their calculations.
- 24:05And the reason why this legislation was passed
- 24:07is that Court experts typically rely on the Bureau
- 24:10of Labor Statistics current population survey
- 24:13in order to do their analysis, right? I know
- 24:15I've worked with many economists who have done
- 24:18that and I've had many economists on the other
- 24:20side who have done that. In fact, one half of
- 24:23economists said that they would consider race
- 24:25and 92 % said they would consider gender in their
- 24:28calculations when projecting out someone's potential
- 24:31for earnings. But the problem with this is that
- 24:36these tables are a reflection of gender pay gaps
- 24:39and of workforce discrimination. So it's automatically
- 24:43undervaluing the damages of women and minorities
- 24:47by hundreds of thousands of dollars. So the consequence
- 24:51of using biased data is that you're using averages
- 24:55that represent generations of discriminatory
- 24:58practices. So this is just perpetuating systemic
- 25:01inequality. It's bigger than just individual
- 25:06plaintiffs as well. So when we talk about individual
- 25:09plaintiffs, I have this example on the screen.
- 25:11So you're comparing a male and female plaintiff
- 25:16who are the same age, and have a substantially
- 25:22different projected earnings. So her and the
- 25:26female, the Hispanic female has damages reduced
- 25:29by $27 ,000 a year over a 40 year lifetime. It's
- 25:34a difference of a million dollars. So that's
- 25:37pretty significant. And another example is claims
- 25:42involving a black female versus a white male.
- 25:46a difference of over a million dollars. So these
- 25:49are substantial differences that are based on
- 25:51purely biased information that is unfairly being
- 25:56applied to these clients. It's bigger than just
- 26:01these individual plaintiffs, though, right? So
- 26:03imagine if a waste management company is looking
- 26:05for a site to build somewhere to dispose of hazardous
- 26:09materials, and the company and local officials
- 26:11have to evaluate the risk that an accident may
- 26:15cause to the surrounding community and consider
- 26:18their potential liability if that happens, right?
- 26:21So if race -based tables are used to estimate
- 26:24liability, there is a financial incentive to
- 26:27these companies and to these local governments
- 26:29to locate the facility among residents who would
- 26:32be more likely to be awarded less money. So inspired
- 26:36by California, Oregon actually recently passed
- 26:39a similar law and I'm hopeful that this will
- 26:42spread like wildfire and this prohibition against
- 26:45considering biased data will apply all across
- 26:50the United States. Another way that we see the
- 26:55bias data applied to our clients' cases is in
- 27:00pain, right? So you have a client that comes
- 27:02to you and they say, I've had this injury, I'm
- 27:05experiencing this kind of pain. And it can seem
- 27:08very straightforward, but when you have an expert
- 27:11on the other side who may be discrediting what
- 27:15your client is saying, it's important to examine
- 27:18the information that they're relying upon to
- 27:20say that your client's maybe not telling the
- 27:22truth, your client is maybe exaggerating. And
- 27:25so it's really, I think, vital for us as attorneys
- 27:28to critically analyze the information that we
- 27:32are relying upon when we're hearing our clients'
- 27:34stories at intake, when we're prepping them for
- 27:37deposition, and also when we're evaluating the
- 27:40opposing expert's opinions and testimony at deposition
- 27:44and trial. Women's pain historically and consistently
- 27:50has been underestimated and men's pain has been
- 27:54consistently overestimated. And I want to be
- 27:56clear that I'm not coming on here as someone
- 27:59who is saying that men are terrible and women
- 28:01are the best or wholly aggrieved, but this is
- 28:05what the data shows. This is something that has
- 28:08been consistently found by researchers and it
- 28:10is not just my opinion. So one of the studies
- 28:14that was done compared a patient's pain rating
- 28:18with an observer's rating. So the patient would
- 28:20come in and say, I'm having a seven out of 10
- 28:23pain. And they would have an observer who would
- 28:26observe the patient throughout the visit to see
- 28:29what did the patient appear to be in pain? Did
- 28:31the patient grimace when they were doing certain
- 28:33tests or when their body did certain things that
- 28:35were consistent with what they were reporting?
- 28:38And throughout the study, they found that women's
- 28:41pain was consistently under estimated by the
- 28:43observer, and a man's pain was consistently overestimated.
- 28:49So that's not surprising given that women's pain
- 28:52is often disregarded and minimized by health
- 28:55care providers. Health care providers often,
- 28:59based on how they are trained or the information
- 29:01they're given in medical school, view women's
- 29:04pain as an emotional exaggeration or are very
- 29:07quick to blame physical pain on stress. And this
- 29:13has led to a pain gap in which women with true
- 29:16medical emergencies are often pushed aside. Now,
- 29:19it's even worse if you are a Black woman, right?
- 29:22Studies show that Black women or doctors do not
- 29:25believe that Black women are in as much pain
- 29:29as they complain about. because of implicit biases
- 29:32that have existed for decades and generations
- 29:36about how black people process pain. And this
- 29:39is a dynamic that stems from slavery during which
- 29:42it was a common belief that black people had
- 29:44a higher pain tolerance than their white counterparts,
- 29:48which is just not any and based in any medicine
- 29:50whatsoever. One area that you see very starkly
- 29:56the difference between consideration of male
- 29:58and female pain is as it relates to heart attacks.
- 30:02So there was a 2000 study that was published
- 30:05in the New England Journal of Medicine that found
- 30:08that women were seven times more likely than
- 30:11men to be misdiagnosed and discharged in the
- 30:14middle of having a heart attack. And why is that?
- 30:17Well, because most medical concepts of diseases
- 30:23are based on the male physiology. There are understanding
- 30:27of the male physiology. And women and men actually
- 30:31have very different symptoms when it comes to
- 30:33a heart attack. Men, typically you see they're
- 30:36clutching their chest and they've got their arm.
- 30:39Whereas women, the number one symptom of a heart
- 30:42attack is actually nausea. And it's very easy
- 30:45for a doctor to attribute, especially in the
- 30:47case of a woman, nausea to something else, to
- 30:50her period, to pregnancy. I think if you're a
- 30:55woman on here, this... call, you've probably
- 30:58been to a doctor's visit for something totally
- 31:01unrelated to your reproductive organs, but still
- 31:04been asked to take a pregnancy test. So I think
- 31:07the problem is that many of our physicians are
- 31:09being trained based solely on the male physiology.
- 31:12And so if nausea, for instance, is not a prime
- 31:16symptom that appears in men to recognize a heart
- 31:20attack, that's not something that physicians
- 31:21are going to be trained to look for in women.
- 31:25Additionally, chronic pain is somewhere where
- 31:27we see a really disparate treatment of men and
- 31:30women. If you have a client, I know we have many
- 31:33that come to us complaining of chronic pain,
- 31:35and it's gonna be a very different experience
- 31:37from a man to a woman. 70 % of people who experience
- 31:42chronic pain are women, right? So they disproportionately
- 31:45have chronic pain problems compared to men. And
- 31:49yet 80 % of studies about chronic pain are performed
- 31:53on male mice or human males. So only 20 % of
- 31:59these studies are done for women, about women,
- 32:02and 70 % of the sufferers of this condition are
- 32:04women. So one of the few studies to research
- 32:08the gender difference in pain actually found
- 32:12that women tend to feel it more of the time and
- 32:15more intensely than men. So again, something
- 32:18to consider when you have a client that is approaching
- 32:21you with complaints and pain or if you are on
- 32:24the other side and the plaintiff is a woman,
- 32:27there are different considerations that you should
- 32:29make when evaluating how their pain scores may
- 32:33be being reported by a treater or what they may
- 32:36be saying to an expert. Another area that I think
- 32:42it's really important to recognize the bias that
- 32:45underlies our client's experience is their experience
- 32:48going to the emergency room. We want to believe
- 32:51that when someone goes to emergency room, it
- 32:53doesn't matter what their race, what their gender,
- 32:56what their age is, they will be treated all the
- 32:59same because of the Hippocratic oath. They will
- 33:01get the great care and treatment that they need.
- 33:03But the reality is that this is not the case
- 33:06at all. So here are just a handful of the ways
- 33:11in which people are treated differently. And
- 33:14again, there are many studies and much data to
- 33:17support these statements, but ways in which people
- 33:20are treated different in their... emergency room
- 33:24experiences. So CT scans, for instance, Latino
- 33:27patients are offered CT scans at a much lower
- 33:30rate than their white counterparts. For pain
- 33:33medication, Black and Latino patients are less
- 33:36likely to receive pain medication at all. So
- 33:40if you have a client who comes to you and says
- 33:42that they went to the emergency room and they
- 33:44were released without pain medication, it doesn't...
- 33:46mean that they didn't need pain medication or
- 33:49their injury didn't warrant being medicated.
- 33:51It just may mean that some systemic bias that
- 33:54exists with how the emergency room is being operated
- 33:57or how the providers were trained caused them
- 33:59to perhaps minimize the pain that this patient
- 34:02was feeling and not prescribe it. Wait times.
- 34:06black patients experience significantly longer
- 34:09wait times than other groups of people. And so
- 34:12again, if you have a client who went to the emergency
- 34:14room for an injury and left without being treated,
- 34:18I would not make the assumption that that means
- 34:20that they were not seriously injured. It may
- 34:22just be that they were there for hours and they
- 34:24were never seen and got frustrated and left and
- 34:27maybe went to a local urgent care or somewhere
- 34:29where they thought that they could get treatment
- 34:31more quickly. Chest pain. Similarly to the disparate
- 34:38treatment for women with heart attack, it's unsurprising
- 34:40then that women with chest pain wait 25 % longer
- 34:44to be seen than men with chest pain. Biological
- 34:49differences in male and females. They did a study
- 34:52of medical students and the study found, and
- 34:55I think this was fairly recently, that white
- 34:58medical students held at least one false belief
- 35:01about biological differences between blacks and
- 35:03whites. And that's a problem if these are the
- 35:05people who are going out in the world and becoming
- 35:07doctors and they hold an inaccurate view on how
- 35:12a black person and a white person may. for instance,
- 35:16experience pain, that is an issue. And they are
- 35:18going to go out in the world and they're going
- 35:19to be practicing and treating people. And they're
- 35:21also going to be training other doctors. Abdominal
- 35:24pain. Women with abdominal pain were 25 % less
- 35:29likely to be prescribed pain medication for that.
- 35:32And that also can come from a bias that women's
- 35:35abdominal pain is period related or stress related
- 35:40versus consideration that it might be a burst
- 35:45appendix or a problem with their gallbladder
- 35:47or something like that. So I think it's really
- 35:50important to recognize that just because an emergency
- 35:52room did not do a particular course of treatment
- 35:54or did not respond to your client in a certain
- 35:57way does not negate their symptoms or injuries.
- 36:00It could be in fact due to these systemic differences.
- 36:05You may discover that your client was never actually
- 36:07seen, that they waited too long. And it doesn't
- 36:10mean that they didn't. experience a serious injury.
- 36:14It just means that they experienced unequal care.
- 36:17And that should be addressed in how you approach
- 36:19the case, how you approach discovery, how you
- 36:22approach cross examination. I think it's really
- 36:26important to consider these real life experiences
- 36:29that your client has had. A very specific area
- 36:35where we see big differences and that I think
- 36:37the largely are in gender is how patients experience
- 36:42and resolve from traumatic brain injuries. So.
- 36:47If you are not familiar with the TRAK TBI group
- 36:51and you are someone who does TBI cases or is
- 36:55interested in doing TBI cases, I highly recommend
- 36:57you read the studies that they have put out.
- 37:00But they did a study in 2000 that studied mild
- 37:04traumatic brain injury patients for two months
- 37:07after injury. And there's the severity and duration
- 37:12of cognitive and somatic symptoms in women was
- 37:18significantly worse than it was in men compared
- 37:22to an orthopedic trauma control group. And so
- 37:26we need to understand, and researchers don't
- 37:28necessarily know the answer for this yet, but
- 37:30they need to understand whether these differences
- 37:34in how women and men process and respond to traumatic
- 37:41brain injury are unique to mild traumatic brain
- 37:45injury or that it is generally the case for injuries
- 37:50for men and women. Because one thing that is
- 37:53very unique about brain injuries is that they
- 37:56rely very heavily on subjective reporting from
- 37:59a patient, right? Women are much more likely
- 38:02to seek out medical treatment. They are much
- 38:05more likely to be more narrative in their complaints
- 38:08about what's going on with them. Men are often
- 38:13taught not to go to the doctor or not to... seek
- 38:18out treatment about things that they may view
- 38:20as emotional, right? Many of the symptoms of
- 38:23a brain injury are not that you can't move your
- 38:26arms or your legs or you're having pain, although
- 38:28pain certainly can be associated with it, but
- 38:30it's often things like I'm having trouble remembering
- 38:33things. I'm depressed. I feel super anxious or
- 38:37I'm very quick to anger. And those things are
- 38:40often and society teaches men not to reach out
- 38:43for help for those. So it's important for researchers
- 38:46to discern and TBI is an area that has only in
- 38:50the last decade and a half really gotten a ton
- 38:53of funding for research, but it's something that
- 38:55researchers need to discern what the differences
- 38:58is. Are they actual biologic responses that the
- 39:02body is having that causes different symptomology
- 39:06or is it a difference in how our society functions,
- 39:11how our medicine, medical providers function.
- 39:15And so that's just an area that we need more
- 39:17research. But the studies that we have done do
- 39:21show that 60 % of those studies show that TBI
- 39:28sufferers who are women experience slower resolution
- 39:31of their post concussive symptoms. They experience
- 39:35more emotional symptoms and they generally have
- 39:37worse outcomes. So again, something to consider
- 39:40asking a defense expert if they're saying your
- 39:42client is fine and should be fine and they should
- 39:45have resolved within a month. Did they consider
- 39:48the patient's gender in making their analysis?
- 39:53And one of the reasons why we don't know so much
- 39:57about how women suffer from TBIs and how they
- 40:03recover from TBIs is that, as I talked about
- 40:05earlier, women have typically been underrepresented
- 40:08in studies and clinical trials. And they've often
- 40:11been... wholly excluded from preclinical studies.
- 40:14And we know that women and men's bodies are not
- 40:17identical. Their bodies, we know, react to trauma
- 40:20differently and they heal differently, not even
- 40:23talking about brain injuries, just talking about
- 40:24injuries as a whole. And so we also know that
- 40:28women are in certain populations that are more
- 40:31likely to experience a TBI. So they're more likely
- 40:34to be at higher risk than men are. intimate partner
- 40:39violence, domestic violence, affects 6 .2 million
- 40:43women per year. And so it is very common for
- 40:48women to experience brain injuries as a result
- 40:51of partner violence. So if women are not represented
- 40:55in these studies, if they... are not being considered
- 40:58when we're trying to find out how brain injuries
- 41:00work and what the resolution for them looks like,
- 41:04then the conclusion from these studies may not
- 41:06apply to women. So if a study says that the majority
- 41:09of TBI patients recover within three months,
- 41:11but that study doesn't include any women, how
- 41:15can that study be used to form an opinion about
- 41:18the recovery trajectory of your client? Another
- 41:23thing that I think it came out last year, a study
- 41:27came out last year is that pregnancy brain is
- 41:29real and why is that relevant to TBI? So I wanna
- 41:34connect the dots for you. A study was done where
- 41:38MRI scans were taken every few weeks from before
- 41:41a woman was pregnant until two years after she
- 41:44gave birth. And what it revealed is widespread
- 41:47reorganization of her brain during pregnancy.
- 41:51And some of those changes were short -lived and
- 41:53some of those lasted a long time. So the most
- 41:56apparent change was a steady decrease in gray
- 41:59matter in the brain in both volume and thickness
- 42:03throughout pregnancy and a temporary peak in
- 42:07neural connectivity. which is like the white
- 42:10matter microstructure of the brain, at the end
- 42:13of the second trimester. And both the cerebrospinal
- 42:17fluid and their ventricles, which are the brain
- 42:20cavities, expanded during pregnancy. And these
- 42:24changes were linked to their rising hormone levels,
- 42:27which is unsurprising. And some of these changes...
- 42:30persisted past pregnancy, right? They were found
- 42:34in the two years post -birth. So that global
- 42:38reduction in gray matter volume was something
- 42:41that persisted. And so this is, again, something
- 42:44that needs to be studied more. who are pregnant,
- 42:47do they rebound? Does their brain rebound to
- 42:50where it was pre -pregnancy or are these changes
- 42:52in gray matter and neural connectivity, are those
- 42:56permanent? Because this study ended after two
- 42:58years. So again, something to consider if you
- 43:00have a client who has a brain injury that either
- 43:03was pregnant at the time or has ever been pregnant
- 43:08and how that pregnancy may have affected the
- 43:11actual organization and structure of that person's
- 43:14brain. Another area that we see bias data influencing
- 43:21how people are treated in both our cases and
- 43:24in their day -to -day treatment with medical
- 43:26providers are in spinal cord injuries. So two
- 43:30parts of the spine, the cervical spine up here
- 43:33by your neck and your lumbar spine down by your
- 43:36pelvis. So the lumbar spine in women is more
- 43:40curved and it makes the female body more susceptible
- 43:45to fracture and dislocation. And it is more curved
- 43:49to support the uterus during pregnancy. That's
- 43:51not surprising, but. it just makes it more stable,
- 43:58but less mobile. So in a crash or something,
- 44:01it's more likely to be injured. And then the
- 44:07cervical spine is more flexible in women. women
- 44:14have a more increased range of motion. So they
- 44:17are typically more protected in their cervical
- 44:20spine area than men and are less likely to have
- 44:23fractures compared to a man. So again, something
- 44:26to consider if you have a client who is presenting
- 44:28with a cervical or lumbar issue. My personal
- 44:33favorite area where we see biostatic influencing
- 44:38our day -to -day as people who navigate the world
- 44:41in a car, as well as our clients' experience
- 44:44if we have a case involving a vehicle's automobile
- 44:46safety tests. So auto design is directly... influenced
- 44:53by the results of safety testing. So any bias
- 44:56in the way that a car is tested translates, obviously,
- 44:59into the way that the car is designed and manufactured.
- 45:03Car companies design their cars specifically
- 45:05to pass crash tests so they can market them to
- 45:09consumers as safe. Hey, this has undergone all
- 45:12these crash tests, it's passed them, and this
- 45:14is a very safe car for you to buy for you and
- 45:16your family. So if the tests do not take into
- 45:20account the differences between male and female
- 45:22occupants, then car makers will not make changes
- 45:26to better protect men and women who, as we just
- 45:30discussed, are very different physically. So
- 45:34women are the majority of drivers in the United
- 45:36States. It did not used to be that way, but it
- 45:38is true today. And they are still historically
- 45:41less likely to get into a crash than men. But
- 45:44they are at a higher risk of injury and death
- 45:48in a car crash. They're 73 % more likely to be
- 45:52injured and they are 17 % more likely to die
- 45:55than met. And why is that? Well, crashes occur
- 45:59in three phases. So there's first the vehicle
- 46:02crash where the impact of the car into a foreign
- 46:05object like a wall or another car or a tree.
- 46:07Then there's the human crash where the body of
- 46:10the vehicle's occupants are coming into contact
- 46:12with things inside the cars, like windshields,
- 46:16windows, seat belts, airbags. And lastly, there's
- 46:19the internal crash, which is the collision of
- 46:22organs, bones, soft tissues within the human
- 46:25body. So crash energy that is not absorbed. by
- 46:30the vehicle in the first stage is transferred
- 46:32to the occupants during the human crash and the
- 46:36internal crash. So vehicle restraint systems,
- 46:40seat belts, airbags are intended to limit motion
- 46:42and limit the transfer of energy, right? That's
- 46:46the whole purpose of them. So to do that effectively,
- 46:50car makers need to consider not just the size
- 46:53differences of occupants, but also the material
- 46:57properties of their bodies, the differences in
- 47:00bone density, the difference in percentage of
- 47:04muscle mass and things like that. It's a very
- 47:08important factor that car manufacturers should
- 47:10be considering, but because car testing is not
- 47:13requiring them to do so, only really one auto
- 47:17manufacturer is doing that. And I will tell you
- 47:19who in a second. So crash tests today, are based
- 47:25exclusively on the male dummy. So crash house
- 47:29dummies were invented in 1976. They did not create
- 47:32one. The government did not create a female one,
- 47:34I should say, because one, they didn't have enough
- 47:37money. And two, men at the time were the ones
- 47:40mostly dying in cars. Women were in the workforce
- 47:43less and they were driving less. So at the time
- 47:46this dummy was created, he was a 171 pound, 5
- 47:51'9 dummy. Men today are actually on average 26
- 47:55pounds heavier. So a female dummy was not introduced
- 47:59until 2003. So 27 years of car safety policy
- 48:05and research was based on the 50th percentile
- 48:081970s male body was not updated in any way. In
- 48:121980, NHTSA actually proposed updating the male
- 48:16dummy and adding a family of dummies, a woman
- 48:19and a child. But in 1981, Ronald Reagan became
- 48:23president and slash and it says budget and that
- 48:26nothing came of that. So in 2003, and it's a
- 48:29finally put a female dummy in a car. But the
- 48:33problem with that female dummy. as you can see
- 48:35here on the screen, was that it was just a scaled
- 48:38down version of the male dummy with boobs, basically.
- 48:42And it represents only the smallest 5 % of women
- 48:45by 1970 standards, not even by 2003 standards.
- 48:50It's so small that it could be a 12 year old
- 48:52child. And women are five and five and a half
- 48:56inches shorter and 27 pounds lighter on average
- 49:00than the average male. But we were just using
- 49:02basically a 12 year old child dummy. since 2003.
- 49:07And no dummy takes into account the biological
- 49:10differences between male and female bodies, right?
- 49:13As we talked about earlier with the lumbar spine,
- 49:15the lumbar spine is different, the cervical spine
- 49:17is different. A female's pelvis is geometrically
- 49:22different than a male's because it's designed
- 49:24to hold and give birth to a baby. The internal
- 49:28makeup of female bones is different. Women are
- 49:31more prone to osteoporosis and osteopenia. We
- 49:34have different muscle mass. Women are more likely
- 49:37to suffer whiplash injuries than men. And car
- 49:41manufacturers also don't take into account, because
- 49:44car testing does not take into account, how a
- 49:46woman might be positioned differently in a vehicle
- 49:49than a man. A woman may be closer to the steering
- 49:52wheel. A woman, I know that I'm closer to the
- 49:55steering wheel than my husband. I'm five two,
- 49:57he's six two. Our settings in our car are very
- 49:59different. We may wear our seat belts differently.
- 50:02I move mine down, my husband moves his up. And
- 50:05car manufacturers and auto amount of testing
- 50:08does not take that into account. So the first
- 50:12truly representative female crash dummy was not
- 50:16invented until 2022. And it was not invented
- 50:19by the government, but by a group of researchers
- 50:22led by Dr. Astrid Lind from Sweden. And she created
- 50:27a dummy who represents the 50th percentile of
- 50:30women in height. So she's, this dummy is five
- 50:34three and 20th five percentile in weight, 137
- 50:38pounds. It has a bust, it has wider hips, it
- 50:42has narrower shoulders, the stiffness of the
- 50:45neck is different, and she put all of the data
- 50:48from all the testing she did with this dummy
- 50:50on the internet for free and makes it available
- 50:53to any car manufacturer or automotive organization
- 50:56throughout the world that wants to see it. And
- 50:59in May 2024, there was a bipartisan bill introducing
- 51:03the United States to include a female dummy.
- 51:08in crash testing and it was called the She Drives
- 51:13Act. And so hopefully that is something that
- 51:17gets passed and that automotive testing is made
- 51:21standard to have a female dummy included that
- 51:24not only reflects our size, but it reflects how
- 51:27our body is actually put together and where we
- 51:31carry our weight and what makes sense. In a perfect
- 51:33world, we'd also have a pregnant dummy, but you
- 51:36know, I'm not gonna get greedy. So we know that
- 51:40these changes can be made to address these differences.
- 51:45So it was already done as it relates to airbags.
- 51:49And from 1996 to 2000, 180 people were killed
- 51:54by airbags in low speed crashes that should not
- 51:58have been fatal. And of that 180, 118 were children.
- 52:03So the government and auto manufacturers realized
- 52:06that it had to do with airbags being programmed
- 52:08to be powerful enough to keep a 50th percentile
- 52:12man in his seat in a crash even if he was not
- 52:15wearing his seatbelt. So NHTSA recognizing, hey,
- 52:18a lot of children are dying because these airbags
- 52:20are too strong, relaxed the testing requirements
- 52:24for the auto manufacturers. to rapidly redesign
- 52:28airbags using a sled test instead of a more expensive,
- 52:33more time consuming version of a test. And nearly
- 52:36one half of auto manufacturers reduced the power
- 52:39of their airbags between 1997 and 98. So in just
- 52:43in one year. And by 1998, NHTSA required car
- 52:47manufacturers to install advanced airbags that
- 52:50deploy with a force that is proportional to the
- 52:53weight of the occupant in the seat. And what
- 52:55happened was there was a massive drop in fatalities.
- 53:00So the only auto manufacturer that is doing testing
- 53:03with biologically correct dummies is Volvo. And
- 53:09they have what's called an EVA initiative. which
- 53:12shares 40 years of research and testing that
- 53:15they've been doing. They have a whiplash protection
- 53:19system that they've installed in their vehicles,
- 53:22and it reduces the difference of incidence in
- 53:26whiplash injuries between men and women. So recognizing
- 53:29that there is this difference, they created this
- 53:31system. They also designed something called SIPs,
- 53:34Side Impact Protection System, to cover the entire
- 53:38window. So it's protecting... who are generally
- 53:42smaller and sit lower and closer to the steering
- 53:44wheel. And they actually created the first virtual
- 53:48crash, pregnant crash test dummy who they called
- 53:51Linda, who represents a woman in her 36th week
- 53:55of pregnancy. And if you're not familiar with
- 53:57automotive testing, a lot of it is now done with
- 54:00computer modeling. And so they actually designed
- 54:02a pregnant dummy. And the last thing I want to
- 54:08talk about is bias in artificial intelligence,
- 54:10a hot topic I'm sure for many people now. So
- 54:13automated systems are not inherently neutral.
- 54:16They reflect the priorities, preferences, prejudices,
- 54:21I guess you could call it the coded gaze of whoever
- 54:24created them and chose to model their intelligence.
- 54:30In modern AI, Right? Data rules. The software
- 54:34is only as smart as the data that is being used
- 54:37to train it. And so these algorithms, if you're
- 54:40familiar with them, they use. They work by comparing
- 54:43one data model to similar ones from the past,
- 54:48but it takes a healthy dose of human judgment,
- 54:50at least at this junction before we've gone full
- 54:52terminator, to figure out the right context for
- 54:55a particular data model. So engineering teams
- 54:57have a significant hand in defining these models
- 55:00and thereby the resulting technology. So if there
- 55:04is bias in what they're choosing to train these
- 55:07models on, there's going to be bias in their
- 55:09results. So one widely used spatial recognition
- 55:12data set was estimated to be more than 75 % male
- 55:16and more than 80 % white. So you can imagine
- 55:20what the results from that data set are going
- 55:23to be. So some examples that you see where this
- 55:26bias reveals itself in AI is that in 2015, Google's
- 55:30image recognition system labeled black people
- 55:32as gorillas. In 2018, the ACLU tested Amazon's
- 55:38facial recognition software with members of Congress
- 55:44against 25 ,000 public mugshots, and the system
- 55:48generated 28 false matches, and 11 of those 28
- 55:52were misidentified people of color. So almost
- 55:5640%. So if there are many more white men than
- 56:00black women in a system, it will be worse at
- 56:02identifying black women. We know this. And all
- 56:06of the companies that were studied for this facial
- 56:09recognition, AI, recognize male faces at a much
- 56:12higher rate than female faces. So this is just
- 56:17not surprising. We have to do better at making
- 56:22sure that the data that AI is being used to be
- 56:25trained on is truly neutral and non -gendered,
- 56:30non -racist data. And as long as we're not doing
- 56:34that and many of these data sets are not publicly
- 56:37available, so there's no way to check, we are
- 56:40going to be consistently seeing biased results.
- 56:45MIT evaluated some companies actually, and it
- 56:48had error rates of no more than 1 % for facial
- 56:51recognition for light skinned men. But for darker
- 56:54skinned women, the errors went up to 35%. And
- 56:59AI systems, facial recognition systems from all
- 57:02of these companies, from Google, from Amazon,
- 57:05from everywhere, have failed to correctly classify
- 57:09the faces of many famous non -white people like
- 57:13Oprah Winfrey or Michelle Obama or Serena Williams.
- 57:16So again, you're seeing the results of that data
- 57:19bias data. This one is, I think could very well
- 57:26be relevant to an injury case with a loss of
- 57:28earnings or an employment case. But Even robots
- 57:33apparently think that women should only make
- 57:3583 % of men's salaries. So this study I think
- 57:39came out last year found that AI often advises
- 57:42women and minorities to ask for lower salaries
- 57:45than men. And that's because AI polls. from existing
- 57:49data, which often mirrors the biases of the people
- 57:52that created and the systemic inequalities that
- 57:55exist within our world. So if the data shows
- 57:58that women have historically earned less, AI
- 58:01will suggest that they ask for lower salaries.
- 58:05So this is a problem when many companies are
- 58:08relying on AI for recruiting. And this kind of
- 58:12shows that these biases are going to continue
- 58:14going unchecked. And the last thing I want to
- 58:19talk about as it relates to AI and algorithms
- 58:25and bias is automotive. I cannot think of the
- 58:31word, self -driving cars, self -driving cars.
- 58:34So if you do auto products cases or you have
- 58:37any cases that you're considering or defending
- 58:39on behalf of a self -driving car like a Waymo,
- 58:43this is important information to think about.
- 58:45So there was a study out of the Georgia Institute
- 58:47of Technology where it found that people with
- 58:50dark skin are more likely to be hit by a self
- 58:53-driving car than a person with lighter skin.
- 58:55So the detection accuracy of these vehicles is
- 59:01lower by an average of five percentage points
- 59:04for people who have darker skin. And this was
- 59:07held true even when they controlled for the time
- 59:09of day or whether the vehicle had an obstructed
- 59:12view. And the researchers in the study found
- 59:17by doing some further analysis that the bias
- 59:20was probably caused by two things. One, The vehicles
- 59:24were the algorithm that was used for the vehicles
- 59:27was trained using too few examples of dark skinned
- 59:31individuals. And it was given less emphasis on
- 59:35learning from those examples that did exist.
- 59:38This study was not peer reviewed. I'm not gonna
- 59:40sit here and say that this is some peer reviewed
- 59:42study. And it didn't test any object detection
- 59:46models actually being used by self driving cars.
- 59:50And it couldn't use any training data sets that
- 59:52are, as far as we know, actually being used by
- 59:55autonomous vehicle manufacturers. But again,
- 59:59the problem with that is that the companies do
- 1:00:03not make this data publicly available. So these
- 1:00:05researchers had to do the best they could with
- 1:00:07models that have been used and created by academic
- 1:00:11researchers and publicly available data sets
- 1:00:13by people who have at least made them available.
- 1:00:18So I think this is information that's really
- 1:00:20important for the public to understand. It's
- 1:00:22an issue of public safety. What data are these
- 1:00:25vehicles being trained on? And is it data that
- 1:00:29is not unbiased and is going to cause disproportionate
- 1:00:32harm to people who have darker skin and who are
- 1:00:35typically people of color? So that is the bulk
- 1:00:40of my presentation. I hope that it was helpful,
- 1:00:43even if you don't have a case that's dealing
- 1:00:46with this specific issue to consider perhaps
- 1:00:51some ideas and some of the information that you're
- 1:00:55relying upon to do your cases and evaluate your
- 1:01:01client's claims and damages. I have just one
- 1:01:06minute left, I think, or maybe not, but I have
- 1:01:08two questions. Let's see. Are these biases also
- 1:01:14prevalent and equivalent in jury verdicts and
- 1:01:17damage awards and other types of injuries like
- 1:01:20housing and racial discrimination? I would absolutely
- 1:01:23think so. And I think that that goes back to
- 1:01:25jury selection and why jury selection is so important
- 1:01:28to make sure that. you are truly getting a jury
- 1:01:33that is unbiased because bias does not mean that
- 1:01:36a person is malicious or that they're out to
- 1:01:39get your client, whether you're a plaintiff or
- 1:01:41defendant, it just means that they have some
- 1:01:42life experience that has caused them to feel
- 1:01:45or think a certain way about. your client's case
- 1:01:48or who they are as a person. So I think it's
- 1:01:51just really important to examine and demand of
- 1:01:55a judge that you are allowed to be given the
- 1:01:58opportunity to examine all the jurors. Do I provide
- 1:02:02expert services? I do not. In the area of bias,
- 1:02:06I do not. I am just a lowly attorney, but yeah.
- 1:02:11So that was my presentation. Thank you so much
- 1:02:13for being here and I hope it was helpful. This
- 1:02:17concludes today's presentation. Thank you so
- 1:02:19much, Molly, for an excellent presentation. Thank
- 1:02:24you again to our audience. Have a great day,
- 1:02:26everybody.