Latest / CLEs You Actually Want to Hear / The New Standard of Care: Federal Cybersecurity Law and the Lawyer's Duty of Competence
Transcript
- 0:03Hello, everyone. Welcome to today's program,
- 0:07The New Standard of Care, Federal Cybersecurity
- 0:10Law and the Lawyer's Duty of Competence. I'm
- 0:13Alapoli Castro, Director of Education and Professional
- 0:16Development at Beverly Hills Bar Association,
- 0:18and I'm delighted to introduce our speaker today,
- 0:21Maho Castro. Maho is the founder and principal
- 0:24attorney of Castro Land Legal, where she focuses
- 0:27on cybersecurity and AI law in English and Spanish.
- 0:31Maho advises startups and growing companies at
- 0:34the intersection of cybersecurity, data privacy,
- 0:37emerging technology and business law. She also
- 0:40advises clients and regulatory compliance, data
- 0:42breach response, governance and risk management
- 0:45and small business advocacy. We're so excited
- 0:48to hear what she has to share with us today.
- 0:49So I'm going to turn it over to her. Hi, everyone.
- 0:54Good afternoon. I'm going to be sharing my slides
- 0:55here. Let's just make sure that we have those
- 0:58on. Can you please confirm that we are all in
- 1:04check? It feels like. Perfect. Yes, great. Thank
- 1:08you so much. Well, thank you again. Good afternoon.
- 1:11Thank you to the Beverly Hills Bar Association
- 1:13for having me. Again, my name is Majo Castro.
- 1:16I am talking to you all today from Austin, Texas.
- 1:18Really excited. I'm the founder of Managing Attorney
- 1:21at Castro and El Ligo. We focus solely on cybersecurity,
- 1:26law, AI governance, and data privacy. And how
- 1:28I got to do all of this? Well, I did my master's
- 1:31in law at UT Law here in Austin, Texas. And I'm
- 1:34really excited about it. It's been a journey.
- 1:37And I love sharing my knowledge with my colleagues
- 1:40just so we know what the new competence is. I
- 1:43have a really popular article that I would love
- 1:45to share with you all that it was published at
- 1:48the Texas Bar Journal. It's called the newish
- 1:51cybersecurity competence for attorneys. And people
- 1:53have been loving it. I'm going to share it with
- 1:55the with the people organizing the program and
- 1:58that way you can get your hands on it as well.
- 2:01I want everyone to hold on to one sentence for
- 2:04the next 45 to 50 minutes that we're going to
- 2:06be together. I want everybody to know and understand
- 2:10that cybersecurity competence is now a legal
- 2:13competence. It's not adjacent to it, it's not
- 2:16a maybe, it's not a depends on the case. We are
- 2:20now responsible and have a duty. to protect our
- 2:23client's information. And that's all that we're
- 2:25going to be discussing today. And so we can all
- 2:29get on the same page. For years, our profession,
- 2:32of course, have been treating data security as
- 2:34an IT problem. So we thought that by buying the
- 2:37antivirus or getting a third party vendor, we
- 2:41recovered and we were able to move on. That era,
- 2:45sadly, is over. And both federal laws and our
- 2:47rules of professional conduct have been really
- 2:50clear about it. And I'm going to show you, and
- 2:53I'm going to be pointing to different regulations
- 2:54and frameworks that are going to confirm what
- 2:57I'm telling you. For quick orientation, you're
- 3:01going to see that my slides are going to be referring
- 3:03to the ABA model rules and the ABA comments,
- 3:07because that's the shared standard that we all
- 3:10have in common. I'm in Texas, and y 'all are
- 3:13in California. And I know that you have your
- 3:15own rules of professional conduct. numbered a
- 3:18little bit different, but I know that in one
- 3:20way or another, California has adopted the rules
- 3:23that I'm going to be mentioning today. So just
- 3:25make sure that, you know, to point to the right
- 3:27ones whenever you're going over your ethical
- 3:30duties. But everything that I'm going to be sharing
- 3:32is something that we all have in common, and
- 3:34we're going to, at the end, going to get to federal
- 3:36frameworks and to industry international standards
- 3:39as well. So my whole practice here in Texas lives
- 3:48in the intersection. I help firms, law firms,
- 3:51attorneys, and founders hold data without becoming
- 3:54the next breach headline. Well, if they're lucky
- 3:58enough. So I'm not gonna be reading the exact
- 4:01statutes to you. I'm gonna tell you what the
- 4:03rules mean on a Monday morning whenever you're
- 4:05trying to figure out what you're responsible
- 4:07of. So. Here's the baseline. So about 20 years
- 4:13ago, if a paper file walked out of your office
- 4:16in the wrong hands, that was an obvious malpractice
- 4:20claim. So we were clear about that. But today,
- 4:23that same harm will look like maybe an associate
- 4:26clicking on a phishing link that hands a stranger
- 4:29your client list by mistake. maybe your engagement
- 4:33letters and even your pending wire instructions.
- 4:37Another problem that you might be crossing is
- 4:39a third party vendor that it's having AI embedded
- 4:44into all of their systems and then getting a
- 4:47breach and your clients being affected now. So
- 4:51that's the type of responsibility that I want
- 4:53us to get into today. As I mentioned at the beginning
- 4:57of the presentation, cybersecurity stopped being
- 4:59an IT problem, and most attorneys are still treating
- 5:03cybersecurity as an IT issue, and our rules of
- 5:08professional conduct have been saying otherwise.
- 5:11Rule 1 .1, it's competence, and it now includes
- 5:14the technological competence as well. Rule 1
- 5:18.6c refers to confidentiality. It is mentioning
- 5:22that this is an active security obligation and
- 5:24not a passive promise to safeguard our client's
- 5:28data anymore. And rule 5 .3, it's talking about
- 5:35supervision. So if we hand our data to a vendor,
- 5:40we stay responsible for what they do with the
- 5:43data at all times. So the federal frameworks
- 5:47will sit on top of these three ethical responsibilities
- 5:50that we have, and behind all of it, it sits one
- 5:55phrase. We need to have reasonable efforts to
- 5:59protect our client's information at all times.
- 6:03That is a threat that is running through every
- 6:05item today, more so because we have a bunch of
- 6:07vendors embedded into our environment, and we
- 6:10have also those vendors implementing new technology
- 6:14Let's call it AI today, but tomorrow can be called
- 6:17a different thing. But the whole premise of the
- 6:20presentation is that we remain the ones responsible
- 6:23as for ethical responsibilities and federal frameworks
- 6:26of what happens to our clients' data at the end
- 6:30of the day. So today we have a plan. We're gonna
- 6:36be chatting in threes, in ideally 45 to 55 minutes.
- 6:40And if you have any questions, please feel free
- 6:42to be dropping them in the comments or the Q
- 6:45&A section for me. And I'm gonna be reading them
- 6:49as we go. So first we're gonna be chatting about
- 6:53three ABA opinions. So that way we get an idea
- 6:57of what our duty looks like in practice. Second,
- 7:00we're going to be chatting about three model
- 7:02rules that I've mentioned already. That's where
- 7:05our duty is coming from. And we're going to be
- 7:08chatting about three federal frameworks for what
- 7:11some of your clients and you are subject to,
- 7:14depending on the areas that you are practicing.
- 7:16I think I'm also going to throw in there some
- 7:18of the industry standards. So NIST frameworks
- 7:21that we should be paying attention whenever we
- 7:24are thinking on cybersecurity and we're thinking
- 7:27about privacy. So we are ready to dive in. So
- 7:34part one, we have the three ABA opinions. These
- 7:37are going to be leading the broad duties of competence
- 7:41and confidentiality. And basically it's gonna
- 7:44turn them into do's and don'ts for modern legal
- 7:47tech. So what attorneys need to be doing and
- 7:49not, whenever we're talking about client's data,
- 7:52whenever we're talking about data protection,
- 7:54cybersecurity, AI implementation within our law
- 7:57firms as well. So, I'm going to start with formal
- 8:01opinion 477R from 2017. Some people call it the
- 8:08wake -up call. It updated the ABA's rules, all
- 8:121999 guidance that said that plain unencrypted
- 8:16emails were fine for client communication. Now,
- 8:19basically, we're throwing that out the window.
- 8:22And the whole thing is that lawyers now must
- 8:25use reasonable efforts. And you're gonna be hearing
- 8:27this phrase a lot through my presentation, reasonable
- 8:31efforts. What are even reasonable efforts? To
- 8:34me, and I'm gonna repeat myself, it's to document.
- 8:38So if you're gonna make a decision, this is a
- 8:39question that I've been getting a lot here in
- 8:42Texas. So I'm part of the technology and computer.
- 8:46commission here for the Texas Bar Association.
- 8:49So we get a lot of questions from our colleagues
- 8:52about technology, technology usage, AI implementation.
- 8:56And the question always is what reasonable efforts
- 9:01mean. So I'm gonna tell you what I tell them
- 9:03all the time. Document. If you're gonna make
- 9:05the implementation of a new tool, you need to
- 9:08compare it with something similar that's out
- 9:11there, if it is that we have something similar.
- 9:14So for example, if you're making a decision to
- 9:16use a new CRM system, let's talk about My Case
- 9:21and Smokeball, I think that's another one. If
- 9:23you're making a decision between them two, try
- 9:25to document why you went with one versus another.
- 9:29Literally just a short paragraph. can save you
- 9:32thousands of dollars if we have a cybersecurity
- 9:34or a data breach issue down the line. So the
- 9:40holding back to 477R, the holding was that lawyers
- 9:43must use reasonable efforts to protect electronic
- 9:47client communications. And because the reasonableness
- 9:53fact is fact -specific, there is no single rule
- 9:57that's going to cover what reasonable means.
- 9:58And that's going to be a whole conversation.
- 10:01And I could chat about it for days. But in practice,
- 10:04you cannot treat all your emails the same way.
- 10:07And we know this. So a scheduling email or a
- 10:10draft of a public filing, that's standard email.
- 10:13It's perfectly fine. or then we can have a draft
- 10:17settlement loaded with medical records, trade
- 10:20secrets or account numbers. Well, plain email
- 10:23could fail this standard. We're not being reasonable
- 10:26about the type of communication that we're using
- 10:28in that specific scenario. Now... We're looking
- 10:32at it at encryptions, secure portals, and maybe
- 10:36another protected channel. So for my practice,
- 10:39for example, if I'm going to share medical information,
- 10:42I am going to prompt my client to use my case
- 10:44portal and upload the medical information that
- 10:48we need through the CRM portal versus a regular
- 10:53email. Ideally, and honestly, in practice, I
- 10:56have my clients using my case portal for everything,
- 10:59unless they're not tech savvy, then we'll figure
- 11:02out how to help them. So the formal opinion that
- 11:06we're chatting gives us the balancing test, how
- 11:09sensitive the data is, how likely the interception
- 11:12is without safeguards, how much the safeguards
- 11:15will cost if you have to put them in place, and
- 11:18how hard are those safeguards easy to use for
- 11:22the client. So those are the type of questions
- 11:25that we should be having in mind whenever we're
- 11:27thinking about client communication and protected
- 11:30information. So ideally, we're building a tiering
- 11:34system. So we're not deciding on email by email.
- 11:37So tier one, for example, could be routine admin,
- 11:41standard emails. Tier two might be client matter
- 11:45updates. We can use a secure portal. And tier
- 11:48three can be high risk, you know, M &A, protected
- 11:52health information, you know, or personal data.
- 11:56Well, in that tier, we should be using encrypted
- 11:58and behind multifactor authentication, maybe.
- 12:01So, and again, it all depends. And the biggest
- 12:04mistake that I'm seeing is that lawyers assume
- 12:06that the email provider's default settings are
- 12:09legally enough and they're not so if we ever
- 12:12face a complaint after an incident the question
- 12:15won't be if we were hacked it'll be well were
- 12:18you making and taking reasonable efforts and
- 12:21documented your decisions for you to have avoided
- 12:24this scenario because even if and it's technically
- 12:27it's going to be impossible for us to get hacked
- 12:30everyone's going to have it at one point or another
- 12:32the question is if you used and implemented reasonable
- 12:36efforts and you documented your decisions. So
- 12:47477R, it's about prevention. And now we're going
- 12:51to be moving on to opinion 483 from 2018. And,
- 12:57well, this is going to be applicable the morning
- 12:59that your prevention methods fail. So there are
- 13:03four duties that are going to trigger it. are
- 13:06going to be triggered the moment that we discover
- 13:08a breach within our system. So one, we're going
- 13:11to have to monitor. We have to be able to detect
- 13:14a breach. I didn't know it's not going to be
- 13:17a defense if we never set up the monitoring in
- 13:19the first place. That is going to be ideally
- 13:21handled by IT. If you're a growing law firm like
- 13:25I am, I have someone that I have a third party
- 13:29vendor. I already vetted them. I documented my
- 13:31decision. as to why I chose them over someone
- 13:35else or another group. So that's done. Two, we
- 13:39need to stop and restore. So the second that
- 13:42we detect the breach, we have to contain it,
- 13:45and we need to bring the systems back. In practice,
- 13:48that means, well, calling the outside incident
- 13:51response firm and breach counsel. And this is
- 13:55going to be the first time that I'm going to
- 13:57mention that if you do not have a cybersecurity
- 13:59insurance in place, It is my best recommendation
- 14:02for you to start to browse in for one. Ideally,
- 14:06you have about $1 to $3 million in coverage.
- 14:10And ideally, you are selecting your breach response
- 14:12council and your forensics council before the
- 14:16breach happens. Right now, I've been working
- 14:18on an incident response plan for a technology
- 14:20company. One of my biggest clients here, we've
- 14:23been working on their incident response plan
- 14:25for about... Two months at this point and every
- 14:29time that I think that we're almost ready There's
- 14:33a gap there is someone that we haven't defined
- 14:35yet There's a you know tabletop exercise that
- 14:37we have to execute to make sure that the plan
- 14:40is actionable So yes, these things are going
- 14:43to take time but believe me when I tell you that
- 14:46while you're working on the implementation of
- 14:51monitoring, selecting your vendors, drafting
- 14:54your incident response plan, you're gonna be
- 14:56working and understanding where your gaps are
- 14:59laying. Three, we need to determine what happened.
- 15:03So we need to investigate the scope of the breach
- 15:05when we have it. We need to try to determine,
- 15:09not try, we have to determine what systems, what
- 15:12files, who's data. was affected. Without that
- 15:17information, we cannot even meet our notifications
- 15:20duty. And that is ideally why we're working on
- 15:23an incident response plan, that it's already
- 15:26telling us what our duties for notifications
- 15:29are. My client has a 45, I think it's a 52 page
- 15:34document by now, in where the exhibit pages have
- 15:38already the notification templates. So we don't
- 15:41have to be thinking on the spot in the middle
- 15:43of a breach. you know, what to say to the clients
- 15:46whenever we do decide that it's time to communicate
- 15:48with them, if we get to that point. And four,
- 15:52we need to notify under rule 1 .4 whenever we
- 15:55have an independent ethical duty to tell all
- 15:58affected current clients when their information
- 16:01is material or materially, you know, compromised.
- 16:06Then again, we have to determine within our own
- 16:08law firms when we're going to trigger rule 1
- 16:12.4, because if I have 5 ,000 clients and only
- 16:16five of them were affected, then we have to have
- 16:19different triggers. Who's gonna notify what,
- 16:22what type of communication are we gonna tell
- 16:23them? Ideally, you're letting the client know,
- 16:27hey, this happened, but these were the remediation
- 16:29measures taken. So you don't have to worry about
- 16:32A, B, and C. So we can put them at ease. that
- 16:37ethical duty exists separate from any state breach
- 16:43notification statutes. We can hit a spot where
- 16:46a state statute's threshold is not met. So for
- 16:49example, in Texas, I think I need to notify the
- 16:52AG's office whenever more than 500 clients are
- 16:55affected. But maybe my ethical duty says that
- 16:58my local ethical rules or local Opinions say
- 17:03that I have to notify whenever, I don't know,
- 17:0625 clients are affected. So we still have to
- 17:09notify clients sometimes whenever even the regulatory
- 17:12threshold is not met. And that is why we have
- 17:15to have a good incident response plan in place
- 17:17that is adjusted to you, your practice, your
- 17:20state, the type of law that you are practicing.
- 17:24That way it is an actionable plan. Every time
- 17:26that I do these presentations, I get literally
- 17:29every time it works like a clockwork. I get about
- 17:33five emails of people requesting me an incident
- 17:37response plan template. I can definitely share
- 17:40it. I don't have a problem. But the template
- 17:41that I have here in Texas with my vendors, with
- 17:45my regulations, with my opinions here in the
- 17:47state of Texas are going to be completely different
- 17:50than you have in California than we have in any
- 17:53other state. So just keep that in mind. So the
- 17:57move to You know comply with opinion 484 hundred
- 18:03and eighty three and our ethical rule ethical
- 18:06Rules ideally is to write an incident response
- 18:09plan before we need it And it's really important
- 18:12for us to try to understand that before we need
- 18:15it So three things at least need to be listed
- 18:18on a decent incident response plan who to call
- 18:21first You know cybersecurity carrier breach counsel
- 18:25incident response firm We need to ask ourselves,
- 18:30how do we preserve the forensic evidence? And
- 18:33who approves the client notice? Who is responsible
- 18:35for the client communication? Because with every
- 18:38breach that we get, we also need to be thinking
- 18:42about the reputational damage. To me, that is
- 18:44the most important and most delicate one of them
- 18:46all. We're attorneys, attorneys, accountants,
- 18:49you know, we rely and depend a lot on our reputation.
- 18:53How would you feel if me as a cybersecurity law
- 18:55firm will get breached and you are part of it
- 18:58and I am not even communicating with you, I don't
- 19:00have an internal process for it, that'll be terrible
- 19:03for my reputation, it'll be terrible for my firm,
- 19:06you know, and these are things that now attorneys
- 19:08need to be thinking of because in the past, as
- 19:11I mentioned, we will outsource it to a third
- 19:13party vendor, to an IT technical team, call it
- 19:16a day. but that it's not it anymore. We have
- 19:19to be involved in the internal processes. We
- 19:21are the ones that ideally are building our incident
- 19:24response plan from scratch and with plenty of
- 19:27time. So we not only build the plan, but we put
- 19:30it in motion. We have tabletop exercises. We
- 19:34train our personnel. So there is an associated
- 19:37thing with the drafting of an incident response
- 19:41plan that it's not only, you know, write a 25
- 19:44page document and that's it. And the one that
- 19:51takes over every Q &A that I have and that I've
- 19:54had lately, it's going to be the formal opinion
- 19:56512 that it was, I think that came out in 2024.
- 20:04It relates to Gen. AI. So it invents zero new
- 20:09rules. It basically maps the duties that we already
- 20:12have onto AI. So the six rules that opinion 512
- 20:18is touching is competence. So basically, we need
- 20:21to understand the tool's capabilities and limits.
- 20:26Two, confidentiality. So we need to know how
- 20:30the tool uses the input data. So if we're using
- 20:33any AI or generative AI tool, we need to ask
- 20:38it in the background, one, please do not train
- 20:41your model with it. And two, ideally, we have
- 20:44a DPA in place, a data processing agreement in
- 20:47place. And that is why it's good to have the
- 20:50team or the corporate versions of it versus the
- 20:54free on paid versions. It also touches communication.
- 20:59So basically, disclose AI use when it's material
- 21:03to the client. And honestly, I am big on transparency
- 21:07in my law firm. That's all that I care about
- 21:09as well with my clients. So I already added an
- 21:12AI clause at the beginning of the year, and I
- 21:15let my clients know, hey, by the way, we are
- 21:18using AI to do these determined tasks. whenever
- 21:23potentially with your data, are you OK with that
- 21:25or not? And we allow them to pick and choose.
- 21:28We allow them to be informed, even though that
- 21:31I might not even be using their data for anything
- 21:33material. So that is on the transparency side.
- 21:37Another rule that it touches is candor. We need
- 21:40to verify the outputs. We know that hallucinations
- 21:43are not a defense anymore. I don't know. It's
- 21:46not a defense. We've seen it. It's not new. It's
- 21:49only going to get worse. and we don't wanna be
- 21:53on the next headline, so check and verify, it's
- 21:58always a good practice. It's also touching on
- 22:015 .1 and 5 .3, supervision, we need to train
- 22:05our staff, we need to set a policy, and we need
- 22:08to vet our vendors. One thing that I've been
- 22:10seeing a lot. people not vetting their third
- 22:13party vendors thinking that because you are choosing
- 22:15a third party company, their responsibility is
- 22:18on them. And not really. Now I see a lot of security
- 22:22questionnaires floating back and forth. Even
- 22:25I had a client asking me for my security measures
- 22:28even before, of course, it was a big company.
- 22:30But even before they engaged me, they sent me,
- 22:33I think it was a short questionnaire, it was
- 22:35less than 20 questions, but they wanted to know
- 22:37when you have an incident response plan in place.
- 22:39Two, I think they asked, when did I have my last
- 22:44tabletop exercise on security in? Three, what
- 22:49AI systems are we using? And so forth. And we're
- 22:52going to start seeing this more often. And we
- 22:55need to be ready to ask those questions. So whenever
- 22:58we are talking about, and this is extra that
- 23:01I'm throwing at you, but because I know that
- 23:04lots of firms are implementing AI, ideally you
- 23:07have an AI inventory. You have an inventory,
- 23:11you know who has access to what, you have a policy
- 23:14in place, you have training. That way, whenever
- 23:16the questioner comes, because they will start
- 23:19floating around more often, you are confidently
- 23:21able to answer those questions to your potential
- 23:24clients. Another rule that it's such in, of course,
- 23:28fees. We are not billing for time saved by AI.
- 23:33It was maybe a month ago that I saw someone on
- 23:35LinkedIn with lots of followers say, or basically
- 23:40complain as to why he would have to stop charging
- 23:44for the time that the AI saved if he had to review
- 23:48everything that the AI system did. And that's
- 23:52a two -fold question. One, the AI system is working
- 23:56for you after you gave it the prompt, and two,
- 23:59then you have the supervision time. That is the
- 24:01time that we charge for. But it was used... Was
- 24:05curious to see how people in the comments had
- 24:07a lot of opinions on this so just to clarify
- 24:11Bar associations and opinions have come out saying
- 24:14that we cannot charge for the time that the AI
- 24:17saved us That's just a common ground these days
- 24:20So We're gonna get into the motor the model rules
- 24:26where the duties are coming from and I know again
- 24:29we're I'm chatting with you know with y 'all
- 24:32located in California, but your state has adopted
- 24:35these rules in one way or another. So we saw
- 24:38how this plays out through the opinions. Now
- 24:41I want to anchor what I just mentioned to the
- 24:44rules themselves. And remember, California, I
- 24:47think it has rule 1 .1, 1 .6, and 5 .3. Please
- 24:52double check and verify that they carry the same
- 24:56duties. And I'm going to mention a couple of
- 24:59things that I read about California as we go.
- 25:04So, including California, our Supreme Court adopted
- 25:07this language in... Sorry. Give me just a second.
- 25:18We talk about General EVI, part two. Okay. So,
- 25:22rule point one, comment eight. It's the competence
- 25:25rule. This is a teeth for the modern practice.
- 25:31here in comment eight, added in 2012, and it's
- 25:34now, I think, taken by more than 40 jurisdictions,
- 25:40including California. So our Supreme Court adopted
- 25:43this language in 2021 as comment one to the California
- 25:46rule 1 .1. So it made California, I think, the
- 25:5039th state to put technological competence in
- 25:53writing, and California got there even earlier
- 25:56through other formal opinions that y 'all had.
- 25:59So the language basically says that we have to
- 26:02stay abreast of changes in the law and its practice,
- 26:05including the benefits and risks of relevant
- 26:08technology, and that it's not only applicable
- 26:11to AI. Today we can name it, it's AI. Tomorrow
- 26:14it's going to be a different beast because how
- 26:17we see that everything is quickly changing, believe
- 26:20me, we're going to see technology that we've
- 26:22never seen before. So we need to start implementing
- 26:26policies within our law firms. and training to
- 26:29our personnel that are a little bit more broad
- 26:32than AI itself. Just implemented to new technology.
- 26:36I'm trying to start doing these presentations
- 26:38with a more broad concept. So I talk about implementation
- 26:45of new technology instead of AI. And I'm trying
- 26:48to be a little bit more broad with my policies
- 26:50because of the same reason. So the comment eight
- 26:54paraphrase says that we have to maintain competence.
- 26:57A lawyer should keep abreast of changes in the
- 26:59law and its practice, including the benefit and
- 27:02risk associated with the use of relevant technology.
- 27:05What this means for us is that technological
- 27:08competence is now a baseline and it's not optional
- 27:11anymore. And I know it adds. a lot to our plate.
- 27:15We have already a lot of things to do on the
- 27:18daily, but now we also have to worry about if
- 27:21our paralegal is using AI or not and if they're
- 27:24using it correctly. That's also part of our competence
- 27:27problem. Two, we must understand that the tools,
- 27:30we must understand the tools that we use to deliver
- 27:33the legal services. So ideally, we're not using
- 27:35tools that are not protecting our client information
- 27:38and so forth. Three, if we use vendors, we must
- 27:42competently evaluate that vendor. Ideally, my
- 27:45best advice to everyone is to have a security
- 27:48questionnaire and start implementing it. There's
- 27:51a lot of people and new shiny tools are coming
- 27:53around in the market. Well, even before you plug
- 27:56them in into your system and your environment,
- 27:58in your cloud, in your emails, please, please,
- 28:01at least, ask them the basic security questions
- 28:05and document your decision as to why you went.
- 28:08with those vendors instead of others. And four,
- 28:13failure to keep up is a competence failure and
- 28:15not a personality trait. And another thing that
- 28:17I've been discussing as well, highly. is that
- 28:21at one point in a year or two years from now,
- 28:24because of course, I've heard also attorneys
- 28:26saying that they're not going to implement AI
- 28:28in their law firms. They have their own opinions
- 28:31as to the implementation of AI. And what I've
- 28:33been telling them is that in two years from now,
- 28:36it's going to be a disadvantage for their business,
- 28:39for their practice. Even their clients are going
- 28:42to see a disadvantage in the amount of money
- 28:44that they're paying that attorney. and the amount
- 28:46of time that the attorney takes to, I don't know,
- 28:49let's say, reply to a demand letter. If you could
- 28:52have used AI and you have attorneys in the practice
- 28:54already answering emails in less than two hours
- 28:58and you're still not implementing new technology
- 29:01to make those tasks easy, then they're going
- 29:04to start seeing a disadvantage in their business
- 29:06model and not only from the ethical side of things.
- 29:10Also, we can chat about the ethical disadvantages
- 29:13as well. but I'm talking about the business side
- 29:15as well. And that's something that I've been
- 29:16trying to communicate to my colleagues as much
- 29:21as I can. So we are 30 minutes in and we are
- 29:26right on track for all of this. So the test for
- 29:31Common Aid, Model Rule 1 .1, Common Aid is simple.
- 29:35So if a tool is relevant to how you deliver your
- 29:39legal services, you have a duty to understand
- 29:41the risk that that tool So for example, I am
- 29:46using AI not for the practice of law, I'm using
- 29:51it for the non -billable things that we all do
- 29:53on the daily. So one good example that I have
- 29:56a really, really good metrics on, it's for the
- 30:00creation of proposals. So whenever a client chats
- 30:03with me for 30, 40 minutes, I send them a proposal
- 30:06letter. It's going to be two to three pages long.
- 30:09They're going to understand what I'm proposing,
- 30:11the fee. Flat fees, I'm using flat fees these
- 30:14days more than hourly. And I explain everything
- 30:17to them really simple. In the past, as of, I
- 30:20don't know, February of this year, I was blocking
- 30:23a full hour of my time for every 30 minutes that
- 30:25I had a client meeting or a client consultation.
- 30:29I would spend 30 minutes chatting with the client
- 30:30and the other 30 minutes creating the proposal
- 30:34itself. I'll write it out. I'll be like, hey,
- 30:38this is what we're offering. Of course, we had
- 30:40a template, but I will spend 30 minutes, 20,
- 30:4222 minutes modifying that template to the specific
- 30:47case. Now these days, starting in May, no, actually
- 30:52April of this year, I automated that. So I threw
- 30:56open APIs and an AI person that helps me, an
- 31:01engineer. We created an internal process that
- 31:05every time that I chat with the client through
- 31:07video, I get the notes and the notes are taken
- 31:10onto Claude in the backend. And the system creates
- 31:15the proposal for me within five minutes after
- 31:17that phone call is over. So instead of me spending
- 31:21a full hour working on a proposal, you know,
- 31:25the 30 minute conversation and the 30 minute
- 31:26proposal. it's only taking me about 38 minutes.
- 31:29So I'm saving a good 20 something minutes per
- 31:32client per week. And in a week, I can easily
- 31:34have 10 to 12 client consultations myself. So
- 31:39you can do the math on the amount of time that
- 31:41I'm saving operationally. And at the same time,
- 31:45I also did my due diligence. So I understand
- 31:48the risks of sharing. the internal notes of a
- 31:52phone call, of the initial phone call with the
- 31:54system, but I documented those risks, I documented
- 31:57the amount of time that I'm saving, how much
- 31:59money that cost me, you know, like I did all
- 32:01my math, I did my internal risk assessment, and
- 32:07I decided that it was a good thing for my firm
- 32:08and where I'm at today. And those are the type
- 32:11of things that you have to be doing as per comment
- 32:14eight. So. The most common dodge whenever I talk
- 32:20technology with lawyers is that, hey, I'm not
- 32:23a tech person, but under common want, that's
- 32:27not defensible. It is a competence gap. And if
- 32:30you cannot explain how your client data is protected
- 32:32right now, today, well, we have a problem long
- 32:36before you had any breach. You need to start
- 32:38understanding your systems today. So these days,
- 32:41as attorneys, we need to be working closely with
- 32:44our IT. Teams. So rule 1 .6c, we're talking about
- 32:52confidentiality now. Now this turns it from passive
- 32:55to active. In California, I think it lands through
- 32:58rule 1 .6 together with business and profession
- 33:01code section 6068E. So same destination, actively
- 33:07protect the data, don't just refrain from disclosing
- 33:11it. I think I got a Q &A. When you use AI to
- 33:17listen in on a phone conversation or video, do
- 33:20you advise a client before you start to avoid
- 33:22any issues of privacy violations? 100%. Yes.
- 33:26So actually, one of the best things that I can
- 33:28advise y 'all is not to have the AI note takers
- 33:33to join by default. I have it on the screen,
- 33:36but I verbally tell the client and I see their
- 33:39demeanor. Some people are more comfortable than
- 33:41others. Out of 10 calls, I've had only like one
- 33:44people saying like, no, no AI note takers, no
- 33:47nothing. I want nothing. Actually, I wanted to
- 33:49meet with you in person type of thing. So it
- 33:52doesn't happen often, but the clients will feel
- 33:55more confident. So just don't have the AI note
- 33:57taker joined by default. Have it join after you
- 34:01verbally notify the client ideally, and you're
- 34:03going to be on the safer side of things. But
- 34:07definitely I use AI. This is a question that
- 34:09I get all the time. They're like, hey, do you
- 34:11use AI note takers? Yes, I do for Zoom meetings,
- 34:15for Google meets, but I don't have anything on
- 34:18me that's recording. Another thing that I started
- 34:20seeing a lot here in Texas and I'm completely
- 34:24bothered about, it's the plot. little device.
- 34:27It's a note taker. It's really small. It's like
- 34:30this small. People are putting it on their wrists.
- 34:33They're putting it on their, you know, on their
- 34:35clothes. You cannot even see it. And we're a
- 34:37one party, you know, consent state. So they can
- 34:39just be recording anything and everything. And
- 34:42it just I'm bothered by it. I find it disrespectful.
- 34:45I find it, I don't know, unethical. I find all
- 34:47of it just completely. Wrong, but just please
- 34:51know that people are being more sensitive as
- 34:54to being recorded as to being you know without
- 34:56getting being disclosed at least So just be even
- 35:00though that we're gonna see new technology that
- 35:02you know how much a plot can applaud device can
- 35:04help us a lot I mean imagine not having to be
- 35:08like I'm taking notes and forget about you know
- 35:10turning the the system on and off That sounds
- 35:13like a dream, but at the same time, there is
- 35:14a lot of privacy and data privacy concerns involved
- 35:19with these systems or devices. So please be mindful
- 35:22of that. OK, so whenever we're talking about
- 35:26model rule 1 .6, so basically, part of phrase,
- 35:32it says that a lawyer should make, again, reasonable
- 35:34efforts to prevent the... Inadvertent or unauthorized
- 35:39disclosure of or access to information related
- 35:44to the representation of a client. So basically
- 35:47we have to do our due diligence. We have to,
- 35:50we cannot be making blinded decisions anymore,
- 35:53period. We're not, that was taken away. The second
- 35:56that we got, you know, new shiny AI tools and
- 36:00implementation of new technology thrown at us
- 36:03left and right. Reasonable effort factors. So,
- 36:08comment 18. One, we need to take into account
- 36:11the sensitivity of the information that we're
- 36:12protecting. Two, the likelihood of disclosure
- 36:15if no safeguards are employed. And again, we
- 36:19are attorneys, even small law firms are most
- 36:23likely to get breached because of the type of
- 36:25data that we have in our systems. So, we have
- 36:30more likelihood to, you know, to be breached
- 36:33and attacked. Three, we need to take into account
- 36:36the cost of additional safeguards. We also need
- 36:39to be taken into account the difficulty of implementing
- 36:42those safeguards and whether those adversely
- 36:47affect the representation with our clients. So
- 36:49if we make it difficult, more difficult for them
- 36:51to, I don't know, upload documents in our portals
- 36:57and such. So perfect. Questions have been asked.
- 37:03I don't know if anyone has questions, please
- 37:05keep dropping them on the Q &A. I would love
- 37:09to answer any of your questions. So now we get
- 37:12into model rule 5 .3, non -lawyer assistance.
- 37:17And here is where we get into third parties and
- 37:22third party vendors, third party vetting, cloud
- 37:27providers, e -discovery platforms, AI tools.
- 37:31IT contractors, transcription services, billing,
- 37:34CRMs, literally everything that we use is going
- 37:38to be implemented by non -lawyer assistance and
- 37:41covered by model rule 5 .3. Three questions that
- 37:47we have to answer whenever we're talking about
- 37:50vendor due diligence. What do they do with their
- 37:55data? What protections exist? and what happens
- 38:00if they have a breach. So those are the, how
- 38:03do we know the answers to those questions? Sending
- 38:05them a mini security questionnaire. That'll be
- 38:09the first and the best thing to do because one,
- 38:12it demonstrates that you try to take reasonable
- 38:15efforts to protect your client information and
- 38:18to your document then at the same time. So ideally
- 38:22today, tomorrow, on Monday, whenever you're more
- 38:25clear headed. You should be counting the vendors
- 38:28that are touching your client data every day,
- 38:30and you need to be doing an inventory. Case management,
- 38:33document management, e -discovery tools, all
- 38:36the ones that I mentioned. If you cannot answer
- 38:40these questions for every tool, you have a 5
- 38:43.3 problem right now, even before you got breached.
- 38:47Because the problems usually start happening
- 38:49way before, but we didn't even stop. and audited
- 38:54our own systems. So the move is that, again,
- 38:59you need to start working on your vendor inventory
- 39:02spreadsheet, list the vendors, the data that
- 39:06they touch, the contract status, the last security
- 39:09review that you did to them, because that's another
- 39:11important thing. So I was working closely with
- 39:15the law firm here in town, helping them with
- 39:16their security, their incident response plan,
- 39:19their training. And they're like, oh, well, we
- 39:21are not going to need you anymore because you
- 39:23already helped us with the initial implementation.
- 39:26All our vendors have already the security questionnaires.
- 39:29They're audited. They're on the spreadsheet.
- 39:31You train our staff. You gave us the. the tabletop
- 39:36exercises, you train the C -suite and you know
- 39:40the partners on all the security stuff that they
- 39:43should know, your technical team came and audited
- 39:46our security controls. Thank you so much for
- 39:49your service. See you never. The surprise is
- 39:52that you have to be checking your third -party
- 39:55vendors constantly because whatever they tell
- 39:57you on that security questionnaire today or you
- 40:00know if they show you a SOC 2 compliant certificate,
- 40:05That might not hold in time in two years from
- 40:08now. So you have to check them accordingly. So
- 40:12once a year, at least you're sending the security
- 40:14questionnaires, you are rechecking the security
- 40:19certifications that they sent you, and so forth.
- 40:22So it is a work in progress, as I like to call
- 40:25it. That'll be a good project and a good homework
- 40:29if you want to really start working on security
- 40:32and vendor and your model rule 5 .3. Now we're
- 40:39going to get into federal frameworks and industry
- 40:41standards. And again, the federal framework is
- 40:44for you to get a brief idea. There's so many
- 40:48of them. And it all depends on the type of law
- 40:50that you're practicing. So we've covered the
- 40:53bar rules, and now we're going to get onto the
- 40:55federal frameworks, which define what reasonable
- 40:58security looks like in the real commercial world.
- 41:01And then here's the good news for a room in California.
- 41:04These three are federal, so it's going to be
- 41:07applicable as they are and as they stand, and
- 41:09you don't have to do any extra research. Of course,
- 41:13we've all heard about HIPAA. You know, it's a
- 41:17hospital problem that it's not gonna people think
- 41:20that HIPAA is a hospital problem and that will
- 41:22not touch law firms But think about a personal
- 41:25injury law firm. They receive a lot of medical
- 41:28information after an accident Do they have you
- 41:32know, are they HIPAA compliant with whatever
- 41:34security they have in place to protect the client's
- 41:37information so HIPAA can reach three groups,
- 41:41you know a the cover entities the providers the
- 41:44health plans they you know And then we have the
- 41:47business associates. Anyone handling protected
- 41:49health information for a covered entity. That
- 41:53expressly also includes lawyers who represent
- 41:56covered entities or handle matters with personal
- 42:00health information. and subcontractors of those
- 42:04business associates. So again, if you are like
- 42:08my example, if you're a personal injury, medical
- 42:12malpractice, employment, even employment defense,
- 42:15healthcare, M &A, you most likely will have access
- 42:19to the records which contain medical records.
- 42:22So now you're a business associate under federal
- 42:25law and you have now a duty to protect that data.
- 42:32We have now three categories of safeguards that
- 42:36HIPAA, you know, imposes. So it's going to be
- 42:39administrative safeguards. So policies, we're
- 42:41talking about training, we're talking about risk
- 42:43analysis. Then it also requires physical safeguards
- 42:47to the facilities and the workstations. And we're
- 42:51also, it also requires a technical safeguard.
- 42:54So encryption, access and audit locks. Who has
- 42:57access to what we have to be audited and, you
- 43:00know, following that all over the place. And
- 43:02of course, we need the business associate agreements
- 43:05and you're going to be also involved in those.
- 43:08So again, if you represent a cover entity that
- 43:11receives PHI, you are now a business associate
- 43:14and you have a duty under HIPAA to control data
- 43:16as well. The GLBA safeguards rule again. It governs
- 43:26financial institutions under the FTC and financial
- 43:29in nature is broad as we know. Mortgage lenders,
- 43:33I can think about tax, I can think about accountants,
- 43:36I can think about even, I don't know, I had even
- 43:40a franchise, I had a company that sells or provides
- 43:44a franchise fee to, you know, they pay for the
- 43:49franchise fee whenever someone wants to buy a
- 43:52franchise through, I don't know, I think it is,
- 43:55small lend -in programs. So they advance the
- 43:59fee for the franchisor. And I thought, and I'm
- 44:03telling this because I really thought in my brain
- 44:06that because they were advancing the franchise
- 44:08fee to the company that they wanted to buy, or
- 44:12the franchisor, they weren't covered under GLBA,
- 44:16but they are. So even things or scenarios that
- 44:19you think that are not touching GLBA, safeguards
- 44:23maybe are. So I was doing, I was neck deep last
- 44:27week onto these whole things. So even whenever
- 44:30we think that we're not, we are under these federal
- 44:33regulations. So who is covered? Again, financial
- 44:36institutions, mortgage lenders, investment advisors,
- 44:39and I put it in bolded because it's now expressly
- 44:43covered, and companies that the FTC defines as
- 44:46financial in nature. So the GLBA safeguards requires
- 44:51that we have a designated qualified individual
- 44:55overseeing the program, and it's literally by
- 44:58name and last name. We cannot be like, hey, cash
- 45:01and legal, it's going to be overseeing the program.
- 45:03they can say whenever they audit, if they go
- 45:06and do investigations on your WISP program that
- 45:11it's not compliant, that it's not enough. So
- 45:15you need to read an information security program.
- 45:17Basically, that's a version of an incident response
- 45:19plan specifically to financial institutions.
- 45:23These are really fun to generate. I love doing
- 45:25WIS programs. Three, you need risk assessments,
- 45:28you need encryption, you need access controls,
- 45:31you need to multi -factor authentication. And
- 45:34honestly, even law firms would need all of that.
- 45:38I've implemented... you know, a type of risk
- 45:42assessments and access controls one way or another
- 45:44within my law firm, and I recommend you doing
- 45:47the same. And four, it also requires vendor oversight
- 45:51and an incident response plan. I think I have
- 45:54a Q &A. Doesn't matter. Law firm hired outside
- 46:04lawyer to do collection case against former individual
- 46:07client. Can that client claim confidentiality
- 46:10breach from time records provided to collection
- 46:13lawyer without first getting written consent?
- 46:17No. No because you are, there's a rule and I
- 46:20don't have it on top of mine but there is a ethical
- 46:22rule that There are some of our ethical responsibilities
- 46:27are let go whenever we are going after a client
- 46:30for payment or whenever we're called to the stand
- 46:33by a judge, you know, to talk about something
- 46:37that the client was doing because of our representation.
- 46:41So there is something about it, but feel free
- 46:43to send me an email. My email is majo at cash
- 46:46on illegal dot com. I have it top of brain and
- 46:49I'm going to continue thinking about it. But
- 46:51no, I would say no. Okay new in 2024 for the
- 46:58GLBA safeguards rule the FTC must be notified
- 47:01within 30 days of discovering a breach affecting
- 47:03500 or more consumers and Good thing, you know
- 47:07now my question is for everyone in the room Do
- 47:10you have a notification process? Because if you
- 47:14don't then you are having other problems even
- 47:16before you get breached Because whenever you
- 47:19do get breached, you're not gonna even know where
- 47:21to start. You don't want to learn breach law
- 47:25If that's a thing, I don't know how to compound
- 47:28it. You don't want to be scrambling through all
- 47:30of the breach notification responsibilities that
- 47:33we have, you know, through the different bar
- 47:35associations and federal standards and, you know,
- 47:39even high industry standards. At the time that
- 47:42you are, at the moment that you're going through
- 47:43a breach, believe me, it's going to be your,
- 47:46it's overwhelming to say the least. So please
- 47:49be working on your incident response plan and
- 47:51your notification process if you have it. another
- 47:56federal regulation, the de facto federal cybersecurity
- 48:03enforcer, the FTC. Section 5 prohibits unfair
- 48:07or deceptive acts or practices. So the FTC has
- 48:11used this for 20 plus years to enforce reasonable
- 48:14cybersecurity reaching almost every business
- 48:18in commerce regulated or not. Again, we're going
- 48:21to be seeing a lot of reasonable efforts, reasonable
- 48:24cybersecurity. But the good thing is that I do
- 48:26have a good answer for you whenever, if you were
- 48:28to ask me right now, which no one has. Maho.
- 48:31What does reasonable cybersecurity means? Well,
- 48:33I can point you to a couple in high industry
- 48:37standards that I'm going to be chatting right
- 48:39after this deck, this page here, and you're going
- 48:45to have something to look at and something to
- 48:47follow. The reasonableness that we were discussing
- 48:50earlier, that's a trickier one, but we're attorneys
- 48:53and I'm sure that you're going to be able to
- 48:54figure it out. So two enforcement theories are
- 48:58the septic part of things. You said that you
- 49:01would protect their data, but you didn't. So
- 49:03example, a privacy policy on your website promises
- 49:06encryption, but under investigation, they find
- 49:10plain text storage. That's a problem. That's
- 49:12why our privacy policy, even on our websites,
- 49:16need to be really specific as to what we do with
- 49:18the data, how we store it, how we handle it,
- 49:21and what we do with it. So no copy paste. from
- 49:24other websites then we also have the theory the
- 49:28unfair theory so your security was so weak that
- 49:31it harmed your consumers so for example not having
- 49:35basic stuff like multi -factor authentication
- 49:37or patching leading to a foreseeable breach after
- 49:41a company or assistant told you hey please click
- 49:44here to update and you didn't update on time
- 49:47so that that could also be another theory for
- 49:50us to be prosecuted under the FTC Two simple
- 49:55rules never promise more security than you can
- 49:58deliver and never deliver less than what's reasonable
- 50:01So those are the three federal frameworks, you
- 50:04know just to give you an overall idea on how
- 50:08other people's responsibility also leaks into
- 50:11ours But I want to close the loop because I can
- 50:14fill the question What does you know reasonable
- 50:16actually looks like and we don't have to guess
- 50:18that much the industry standards that I was mentioning
- 50:21was the NIST cybersecurity framework from the
- 50:24National Institute of Standards and Technology
- 50:27version, I think 2 .0. It has six functions,
- 50:32basically govern, identify, protect, detect,
- 50:37respond, and recover. A good incident response
- 50:40plan is going to follow those six functions,
- 50:43and I'm giving you already a tip for your incident
- 50:46response plan. So build it around the six functions
- 50:50of the NIST cybersecurity framework. um decided
- 51:00i think i got a question how do you deal with
- 51:02the recent decisions that hold that information
- 51:04is not protected by the attorney campus oh i
- 51:07love this question okay so look i was reading
- 51:10something really interesting at least here in
- 51:12texas i got an opinion we got an opinion i think
- 51:14yesterday or today in where the the judge said
- 51:20that, and I'm gonna put it on my LinkedIn after
- 51:23our chat today. So if you wanna follow me, I'm
- 51:26gonna be posting about it. So the opinion...
- 51:29to paraphrase, basically the judge said, hey,
- 51:32by the way, client, because you were using AI
- 51:35as per your attorney's instructions, you are
- 51:38covered. So that's going to be covered as an
- 51:40attorney -client privilege because you were instructed
- 51:43by your attorney to prep for whatever they ask
- 51:46that person to prep. And then AI is not protected
- 51:49by the attorney -client when you say, I welcome
- 51:51any kind with a client. I think that's going
- 51:53to be, that's being pushed to the side. I do
- 51:55think, and I see that If we're using AI and it's
- 52:00paid version and you have a DPA in place and
- 52:03you're not training their model, you're gonna
- 52:07be good. Those are my three questions to ask.
- 52:11So free versus paid, data processing agreement
- 52:14in place, and no training of models. I think
- 52:17you have a good, not excuse, but a good... you
- 52:24can make a case around that. You can literally
- 52:26say that you were using a system like we use
- 52:28Google and we research, but it's now called Claude,
- 52:31it's now called Harvey or whatever. But I think
- 52:34that judges are moving more towards and leaning
- 52:36more towards the idea that us relying on AI while
- 52:40under those three premises that I mentioned,
- 52:43we're covering it's going to be considered attorney
- 52:45-client privilege. But I love that question because
- 52:48I was just reading that decision today here in
- 52:50Texas. I think we have about eight more minutes.
- 52:55Okay, another industry standard that I will point
- 52:57you to if you're working on your incident response
- 53:00plan, ISO 27001, International Standard for Information
- 53:05Security Management System, certifiable. So when
- 53:09you vet a vendor, you can ask them, are you ISO
- 53:1227001 certified? So even a... Instead of you
- 53:16sending a security questionnaire, you can ask
- 53:18them for whatever certifications they have. And
- 53:21again, remember that you need to be asking those
- 53:23questions every so often because they can lose
- 53:27the certification. And three, the SIS controls
- 53:30from the Center for Internet Security. It's a
- 53:33prioritized list of 18 specific safeguards. This
- 53:37is the one that I use. It's specific. It's ideal
- 53:41for solo or small law firms because we can actually
- 53:43implement it because it's tiered. We start with
- 53:46the basics. It guides you through like, hey,
- 53:49this is MFA, know your device, control access.
- 53:52So that's where I started whenever I opened my
- 53:55law firm and I wanted to start building my security
- 53:58controls. Of course, I have friends and friends
- 54:00that came and checked my security controls and
- 54:02everything. But that's a good starting point,
- 54:05the SIS controls. So take a look at the SIS controls,
- 54:09the NIST cybersecurity framework, and ISO 27001
- 54:13if you want to start working on your incident
- 54:15response plan. Again, when we say reasonable
- 54:19efforts, we are not waving at something that's
- 54:23nonexistent. We have a published bar whenever
- 54:27we're talking about reasonableness and cybersecurity.
- 54:31We don't have to be certified, of course, but
- 54:32we have to know that there are the measuring
- 54:34stick of the industry standard, and most of the
- 54:38regulations come from some version of those standards
- 54:43already. three things to take away with you.
- 54:48One, cybersecurity competence is a legal competence.
- 54:51It's part of what we do, even though for some
- 54:53it's boring to me. I'm passionate about it, as
- 54:55hopefully you were able to tell through this
- 54:58conversation today. I love it. I love not necessarily
- 55:01understanding everything, but always reading
- 55:04and catching opinions and catching the industry
- 55:07standards and chatting with the engineers and
- 55:10understanding this practice way more. Two, we
- 55:13need to have an incident response plan today.
- 55:16And three, we need to understand that reasonable
- 55:19efforts, it's a documented decision. You document
- 55:23everything, please starting today and go back,
- 55:26start auditing your systems. Bernadette, would
- 55:29you recommend having language in agreement with
- 55:31provider? They notify you with 10 days if they
- 55:33lose any certification. 100%. But I will think
- 55:36that 10 days is, depending on the provider. But
- 55:38if you're... It depends. It depends on who the
- 55:42provider is, but I think 10 days is kind of tight
- 55:44because they're going to be trying to figure
- 55:46out and scrambling still, even if you were affected
- 55:50by it. But definitely similar language can be
- 55:54implemented. Thank you again. My name is Maho
- 56:00Castro. I'm located in Austin, Texas, and it
- 56:03was lovely to be, you know, sharing this full
- 56:05hour with you all today and answering your questions.
- 56:08You can find me on LinkedIn. I have a sub stack
- 56:12that is called the Cyber Law Gal. It's in English
- 56:14and Spanish. I chat everything technology. I
- 56:17chat everything privacy and AI. as I know it
- 56:21with my, you know, plain language. And I'm always,
- 56:24I love the question. So if you have any, you
- 56:25have my email here. Actually, it's Majo, M -A
- 56:28-J -O at Castrol and Legal. Always happy to help
- 56:31my colleagues to get through security questions,
- 56:34even cases that you have and where you don't
- 56:36know how to handle some clauses. Always here
- 56:38to help. Okay. So thank you so much again. Thank
- 56:40you for the Beverly Hills Bar Association for
- 56:44having me. And I hopefully we'll see you next
- 56:47time.