Ethical Considerations of AI Usage in Marginalized Communities - Jessica Hoffman & Yoel Alvarez

bi audio yes good morning everyone um I think that the uh audience kind of like split another section going on the other side please excuse the noise and hopefully those attending here can hear as well uh it's going to be an interesting discussion so hopefully we'll have some time for q&as and if anything comes up please just raise your hand and we'll try to answer it as best as possible but uh my name is Joel Alvarez uh intro is coming up here um a cyber security professional with uh just over 15 years of experience um I work for an mssp in the area uh AI taken center of attention for most of the presentations these days

given the impact so uh Jess and I put together a presentation uh that I think has a different take on AI and the impact on society in general but uh we're hoping to open it up for discussion and then make it interesting enough for everybody to interact with us uh as we go through here Jess all right what's up everybody good morning so great to see so many young people here yes okay I'm Jessica Hoffman I work for the city of Philadelphia really love our okay that's what I was waiting for okay okay I'm the deputy siso for the city as well so you know I'm really excited to be talking to you all here because cyber

security is so important to safeguard our data um background and aut compliance and yes I love long walks on the beach God I guess so why are we here what are we talking about ethical considerations for the usage of AI in marginalized communities oh my gosh um so if you're familiar with Philly great City but yes we have some marginalized communities there are communities right here in our city that don't have proper Broadband that don't have I'm losing my hello okay hello the mic's still here okay we don't have proper internet access we are behind in technology hello hello hello so is that me okay all right we have a digital divide we have a digital Gap in

our technology AI is not any different okay so that's what we're going to talk about today how the usage of AI as well as Tech and General has impacted our ability to progress technologically all right let's talk about it hopefully that the change in volume doesn't impact that yall can hear as well here but um it's interesting that as we were talking about how to put together the presentation uh we usually look at this AI World from the technical point of view but we tend to ignore to an extent the social component of cyber security so I was thrilled when she mentioned that this was the idea for the presentation because we get an opportunity to feed on that impact of AI

that has nothing to do with the full development with the tools are going to use it rather the consumers of those products at the end the society so when we look at the challenges that we presented here you're going to get a better idea of the full presentation right and as you mentioned those communities that don't necessarily have full access to information it doesn't mean that they're not going to be impacted by Ai and these are the challenges that we're going to discuss here one is of course access to information and then we'll be looking at the impact of using AI for the tools that will craft the information that this communities get the volume of

information that is generated given the new tools available for AI and again these tools are not dependent on regulations or full development to be available this is the chat gpts of the world already available the image generators the video generators The Voice generators that are available for anyone and everyone to use so the challenges that we're talking about exist today this isn't related

volume oh whatever works best yeah so uh as we look at what can be done now that most of these tools are available today and the level of information is already reaching those communities on how limited fashion that is the way they get the news the way they get the busws the way they get calls or simple media having an impact on their day lives we're talking about someone applying for benefits we're talking about someone doing research on a school or a medical assistance or something related to their day-to-day life the volume of information they get is massaged is controll and to give you more context anyone that has interacted with an online portal today has to face a

chatbot first and then maybe if you're lucky you'll get to a life agent but you have to go through that filter now if someone that doesn't necessarily know how to navigate that technology that brings to enhance the Gap we mentioned already for communities that don't necessarily have that up to date so as we go through the presentation itself we'll talk a little bit about regulations how that has an impact on transparency of the development of these tools the use of these tools the level of Mis misinformation especially around volume of data being posted today and again we're not waiting for this to happen this is already happening the scam the calls the texts the emails the

presence on social media and I'm talking about across Generations across audience right this isn't just for those communities we're talking about this is across the board and of course the level of awareness that we have today and the level of awareness we hope to have to help not only like I said these communities but across the board um I do not know if this is working I'm just like pointing at it we're good thank you but you may Wonder okay how do we get to that level level of visibility into what these tools are doing and the ones being developed will do eventually right and unfortunately we have to rely on larger Scopes what I mean by larger

Scopes is regulations at a government level at a local level at a workplace level but I want to bring it down to a home level because we talk about awareness here and I love it that we have an audience that is looking and seeking information about this because you have to go back and talk to your peers at home talk to grandpa and grandma if they don't have access to any of this or to the kid that is being exposed to that online so we talk about regulation from the highest level and we take a page of what's common for all of those and what we can do at a home level and hopefully spread that information

Beyond home and take it there uh I'm going to briefly cruise through this but I think we can summarize it that the main focus of what is being done in the European Union what is been done here in the US what is been done at a state level the common denominator is about privacy of the data avoid profiling of anyone that has data massaged by AI avoid creating answers to Solutions and requests just off of an automated system and have some human oversight not only on the development but on the decisionmaking process by those AI so love what the European Union did in addition to giving us the USBC port on the iPhones right um they highlighted under regulation of

the bat what is unacceptable period for anything AI related and I love it because we know that anything coming from the EU has a lot of teeth in the development of these tools and we do take a lot of notes from what is implemented on the EU and bring it to our own executive orders that happen at a federal level but also at a local government level for us here we have an example for the one we have here um we can go over the details but it's it's this is just a summary I do encourage people please to go back and at least give it a glance it is very eye opening it is very detailed on what to look for

if you're trying to develop your own security measures to fight what AI could be and to fight what AI is today um and again we're talking about social scoring how do we classify that data that we're massaging this has an impact like we mentioned on those communities today let's talk briefly about what the US is doing and I mean I'm not happy about it we have an executive order that was signed because we had a summit in Europe and we had to go into this Summit with something signed but we didn't have anything because what first went through the house it didn't get anywhere and what did went to Congress is still in Congress we haven't seen anything there

so thankfully something at least was signed I mean it looks like it was written by 10 different people and put together but hey it is something the common denominator is of course privacy data security development oversight and then accountability after the fact if you come and done present the way it's done how do we give that teeth well we'll see that implemented as we make progress on these things right I do want to highlight at a local level Frost and this is just a snapshot from like a week ago or so about a lot of the states are taking action on AI already and you can see the color code the states that have proposed legislation

those that have enacted and those that have enacted and proposed actually but you see a lot of gray and that is the states that is waiting for something at a federal level to do because they haven't done anything at the local level just yet and this is state level uh thank thankfully PA is one of those that does have something again we fall on the same category looking at privacy the use of personal data biometric data how is that being used is it used to profile those that enter the information those that use the tools and then of course it includes language around data privacy for hipop PCI pii which is language around personal information for the most

part but um this is the executive order I want to actually pass it on to Jess she had some say on that and she has of course more context to this which is important for us thank you y so right so as I mentioned I work for the city of Philadelphia but you I live in I live in Pennsylvania and as you just saw by that map every state is a little different when it comes to privacy period uh first of all who cares about their privacy in this room thank you uh because a lot of times especially young people you don't care about your priv it's already out there who cares right um I know you know this

presentation we're talking about artificial intelligence but it's all it's all connected right so I was just actually had a state conference yesterday which was really dope um the the theme of it was uh protecting our data across the board and accessibility usability accessibility um providing value right instead of having a 10-page document that you got to complete to get food stamps or get some kind of Medicare coverage for the state you know let's as you'll mention have like chat Bots right so you can speak to what you need exactly and get rid of all the fluff right so usability of the forms from a state level City level municipality level right we can have benefits and

we'll go into that of AI but what I wanted to share though is that the state of Pennsylvania is actually really far ahead of the game um the department of of Homeland you know they put out uh a lot of money to each state billions of dollars related to cyber security and implementing cyber security so Pennsylvania uh we're like one of the first like three states that responded back with our cyber security strategy plan for that money and what that does is it puts us all on an even playing field for every single County municipality city across the state of Pennsylvania and with that money we're able to invest in the safe and secure usage of AI we're able to invest in

usability of these State applications and local applications that we need especially in the marginalized communities because let's be honest all right in the cities inner cities we have a lot of poverty right and that is a lot attributed to why we don't have the tech we don't have the skills that we need and if you don't know I'm gonna give a pitch real quick go out there in the vendor we got blacks and cyber blacks and black girls hack girls hack women is cyber all of that because we need to make sure that we continue to provide awareness to all these things but at a state level we are actually in a really good place um forward thinking like how

is AI going to impact uh everybody right not just right know we're talking about marginalized Community sh but really AI impacts everybody and we need to be aware of what data is being collected what are they using it for that transparency

issues uh I think that this overlaps with what we're trying to explain in the presentation itself right uh we can talk about the benefits of providing some oversight on this but this is just a highlight and we're going to touch base on a few but of course hopefully this Sparks questions on your side and we can go from there but um oversight at all levels and this isn't just a government issue here this is us making the decision on what tools to use and how do we advise those close TOS that matter either a co-worker or a family member or someone with an audience that has the ability to express his thoughts right oversight it neverly takes to compliance

you have to have a set of requirements for how these things are implemented and how these things are being used but those of us that have had to do a form for anything medical related or School related or benefits related knows about red tape and how many steps you have to take to get to that final decision of a yes or a no so what we're hoping to see is the oversight and the controls ease that process into the decision making but hopefully it does have some human oversight on that step of a yes or a no or an approval or a denial or the next steps to take so the development of those tools has to follow a set of rules

that we're hoping to see on this regulations that have like I said I keep using the word teeth that has the ability to deny a development of a tool says you are not compliant with x y and C you're not allowed to continue we are seeing here industries that have regulations look at Health Care look at overall data processing and we see them bypassing and finding ways to circumvent the regulations constantly we see them misusing this and at the end was that a fine something that has an impact but they continue to do business so it's up to us to create the awareness and not rely on that regulation alone to have an impact that's where we talk about

expedite processing for the information yes response time to get the answer yes automation to get the answer yes but at the same time was all of that process thoroughly documented so if I have a way to challenge it we hope that Automation in general and AI has the ability to catalog all of that and build a file that we can go back and reference if we need to fight it if we need to present it again if we need to say hey you need to reapply for the benefits because the documents just disappeared well we're hoping that this solves a lot of that background to make it happen sooner rather than later of course with that there's going

to be risks just if if you want to talk a bit about this too um this is kind of correlated to the challenges we talked about but again the challenges are more like overall this is more specific to what we see today happening so Jess sure um misinformation disinformation Mal information all those things right uh you can sum it up as bad bad information right so with you probably know like AI can do the voice generation you know picture photo all that stuff so you know imagine somebody who doesn't know about AI doesn't know that that type of Technology exists and they get a phone call at two o'clock in the morning from a grandson you know

needed money and it sounds just like them so that goes back to us talking about spreading awareness taking this back right so just understanding what what AI can do and I mean we honestly don't even know all the things they can do yet right because it's still pretty much in beta um and that's number one reason why a lot of government aren't adopting it fully yet I mean going back to the PA state they're looking at it they have a lot of programs code PA is one of them there's a lot of opportunities for you young folks out there to get into this space because AI is going to generate a lot of new opportunities a lot of new positions and

and on the benefits of this as well but from a risk perspective you know the misinformation is real and that could impact that's like stealing money from somebody that doesn't even have money right and we we we need to be very careful there or stealing benefits going back to that you know being able to to uh we all know about identity theft but that impersonation as well especially with the facial recognition and all that right you can literally copy somebody like that in like seconds right videos content we've been talking about that um some things like around like healthare uh you know that could be compromised as well if we don't have enough information on what the health care is needed for

each of the community you know going back to the fors I mean could you imagine like something simple like a college application or something like let's put it in cont context for us or scholarship application using a chatbot to to be able to streamline that well it would be so much easier and beneficial for those people who aren't techsavvy who don't speak English uh you know that that have restrictions and disadvantages and i' also like to say it's not always just about disadvantage from a knowledge perspective it could be disadvantage from a physical aspect too that's the accessibility or even disadvantages from you don't have the tech or you have crappy Tech or whatever right um so the

risks as far as you know what we can see right now for sure is absolutely privacy and it's awesome to see so many you care about that um I'll let you'll touch on some of the other things but I will say Legacy system perspective kind of ties into the disadvantage of having low Tech we're going to see a lot of outdated technology that just can't keep up with it and that's nothing new for those of you who've been around that is nothing new right we got to accept the risk watch the risk mitigate the risk all that risk stuff um but this is this could potentially further that divide yeah to that context uh those of

us that have been in the industry for a minute know that you always have that one server that one system that lingers it's just you can't get rid of it right so how is that going to be integrated into the next fast solution well we see that as the bottleneck and the security flaw that might have a wider impact than we originally envisioned especially since a lot of these AI capable Solutions are going to be rushed to the market because they don't want to miss out it's the fear of missing out and we see that being displayed everywhere so with that rush we tend to ignore compatibility we tend to ignore integration and the ability to have

those Legacy systems feed information or receive information from so be aware of those and again as much as we try to expedite or want to see expedited the process to get that final answer we see Legacy systems as the one holding that back there will be one that is going to be just saying no waiting for something to respond that should be a second right but um again risks about privacy and confidentiality we talked about it and we love to see the regulation take place on that but best practices and I know Molly you're going to love this because I mentioned vendor risk management and we talked about that recently um awareness and training and I have to

insist on that because at home we try to avoid talk about work right but yet here we are mentioning hey I learned this I'm seeing this implemented I'm seeing this being a reality today so that's the easiest way to translate something into plain English for everybody to understand the moment we mention artificial intelligence they're like huh but if you say hey the volume of information you're seeing now on social media on the news on the way the paper is coming in on the way you're receiving information is definitely being impacted by AI because you can generate it within seconds instead of having to spend a full week developing a short video you can literally have something get a text and

develop a full length of something can be posted also instantly so if you think that I'm trying to massage a subject or something that is important I can have a full room of people developing content about an idea and I can change the perspective of that idea I can actually massage the data to the extent that okay I might not hit the full community but most of it will if I try to do it on a generation I put it on Tik Tok if I try to do it on an older one I put it on Facebook guess what I'm hitting the wide spectrum of generations with the same video the same content that I didn't

generate I was paid to massage data I implemented it and it's out there today this isn't something that is coming down the pike so awareness and training adding the human oversight is also a best practice like I mentioned if you have the ability to at least control how's the how that's being done please do so if you have the ability to minimize the privileges on this AI tool before something is published before something is automatically generated and is out there you stop it you provide the oversight and then you allow it access it's like giving it keys to the kingdom I do have an example of someone enabl in some AI in social media and the thing

just started posting videos by itself it didn't even ask for consent or hey do you approve of this it just follow a profile it builds what this content creator likes doing and it came up thank you I know right but uh the risk management that we have to do is about verifying those sources it's about having to go back and say hey are you aware of this talk to your peers talk to the trusted sources that you have and if you don't have someone to reach out to there's plenty of information available today that you can go in and get but just base your research on those trusted sources and then make a decision don't just rush it

because of the boss word you have to find those sources keep up with the news I mean I know it's too much too little time it's just coming at you like a hurricane right but I think that it makes sense for all of us to at least take the time do the research spread the love and spread the awareness but I'm going to leave you guys with some final thoughts uh from Jess here here and then I'll add some of my own at the end too but AI is here we just have to embrace it we have to be aware that it's here we have to be aware that is impacting us today this isn't coming tomorrow it is having an

impact we don't have to see the next Chad gbt being released to know that you are getting information generated by Ai and again this isn't an antivirus this isn't like a full product this is live and available today to all of us I know we're wrapping up we'll be around if you guys want to talk after but I just want to leave you I'm a huge fan of generative AI I use it all the time I teach a pen State I tell my students to use it yes I do why because there's a time and a place you use the Chad gbt to get your thoughts across or reord something and then you do your research

and fill in the blanks these are tools that we need to embrace not be afraid of right so I challenge you to use them in a safe secure responsible way and that's what we need to be sharing that knowledge as well as the potential abuse of it to those of us that know family members no friends that might not be as techsavvy as you are okay because we are brilliant here you guys are brilliant all right I'm gon I'm gonna say goodbye I'll check you guys later lck in with me thank you thank you thank you thank you uh I'm going to leave you one one last thought although we we focused the presentation around

communities that need the most help uh for anyone here involved in security awareness training at a local level knows that you have audiences within your environment that pay no attention to this so those that are not acknowledging the importance of being aware of these things are as vulnerable as those that have no access to information so when we try to extrapolate this into the work environment into our schools into our homes look at it through that lens is like what's the difference between ignoring the Technologies there and not having access to it what's the difference between not using it at all or ignoring that is there and just like saying well do I really need it and move

on well think of that and use that as a subject to develop your next tool the next time you talk about this the next time you do some research is like who is the most vulnerable who is the most exposed and how can I help thank you so much