Use-Cases for Artificial Intelligence and Machine Learning in Web3 Infra | Jasmine Brunson - Rektify

[Music] artificial networks Etc but these are the things that I've been thinking thinking about and what keeps me up at night and then also what to look out for right uh we're looking at happening 2024 this April and after every happening we see a Bitcoin Bull Run we see a whole bunch of projects propagate but now that AI is hot we're going to see a lot of AI coins so we're going to talk about the differentiation between which AI coins really have utility or might have utility and which ones are just shitcoins and in also one of the most passionate um subjects that I have in AIML in general is ethical and responsible AI ethical Ai and responsible AI so we're going to get into that today and so one of the major applications for AI and web 3 is the Oracle problem issues with data feeds issues with um inaccuracies of the data that oracles get so at the end of the day when we talk about trying to Prime that data making sure that data is accurate so the smart contract actually is working off of accurate data AI does help with that also the Privacy piece in the data Integrity is very important as well and so going back to what I was saying about use cases Dow governance is a major use case right identity making sure that uh we stick to the basis of data science right we have the relationships between one to many uh many to many Etc but you don't want in your Dow for one person to have multiple accounts so they can uh they can vote uh a duplicant amount of time so ensuring that da uh data and identity and identity verification is on point is is very very um needed and necessary and then also with voting mechanisms as well and so funny enough zero NE knowledge artificial neuron networks is a topic that I was pondering one night and I was like why isn't this a thing right why can't we integrate zero knowledge in artificial no networks that will fix the training data problem that we have with blackbox attacks gray box attacks and white box attacks and then instead of going on Twitter and saying I found out or I created zero knowledge artificial know networks I looked it up and I came across an article by a man named Ross alri um some of you all may know Ross albre was a creator of Silk Road he is deling a life sentence in prison for his creation of Silk Road which was a platform that went in straight disarray and to make his right wrong he had talked about the censorship problem when it comes to platforms such as what WhatsApp when it comes to platforms such as Telegram and with that censorship problem zero knowledge artificial know networks can support with that and also when we think about content mod moderation content moderators from Facebook to Twitter Etc they have to go through a lot of Nefarious harmful content whether it be videos whether it be images of graphic content and that disproportionately affects people when it comes to mental health so if we can leverage Z knowledge artificial networks when it comes to the censorship problem on platforms then we can mitigate those problems when it comes to the human level and then also one of the most passionate uh subsets of this space for me is web 3 security how we can integrate AIML to detect smart contract vulnerabilities how about the security researchers the smart contract Auditors so they can focus on the more nefarious bugs and attack attacks and those vulnerability detection tools can help us out with time and money and uh just the future of smart contract creation too often times uh we look at projects like chat GPT and we say hey okay we're going to use chat GPT to create a smart contract well I would say let's hold up on that because a data scientist a nefarious um hacker can embed and do some data injection in that code that uh chat GPT is quering and you might suffer from a black box box attack in the future and so let me just start off by saying this is not Financial advice at all but AI coins are going to be a topic in the next couple of years and I want people to kind of understand that not all AI coins are made the same there are some AI coins that have been propagating before chat GPT launch in 2022 so they might in fact have use case but I just want people especially noobs and people who are coming into this space for financial sovereignty to know that not all AI coins are made the same and some AI coins actually have utility gaming tokens um leveraging AI for tracking onchain data data organ organization prediction models markets ET Etc so going back to what I was saying we're going to see a lot of I coins in the next couple years especially with the Bitcoin Bull Run coming up so I want people to be very very keen and aware that it's going to happen and we have to be smart when it comes to doing your own research and so web three security here are a couple artificial neuron Network um applic ations that I've been researching for the past two to three years um these are applications that can be attributed to fishing dos Etc and so uh Mart to is one where I've researched where that can be used for prevention of Dos on blockchains Tron is a blockchain that had a Dos um problem back around 2014 2015 and dos can render your blockchain useless so we don't want that when it comes to this space self-organizing Maps particularly can be used for dos for l2s graph neural networks degreef free graph convolution neural networks can be used for smart contract vulnerability detection and it also self-structuring neuron networks can be used for preventing fishing so we seen in the nft Run there's been a lot of fishing attacks when we look at web 2 80% of fishing attacks was was distributed through email when we look at web 3 particularly in nfts the metric is similar but it's not through email it's through Discord it's through Twitter so these are some neural network applications that we can use and pivoting from that I just want to say that the Zen of blockchain the Zen of web 3 we talk about access we talk about inclusivity trustless systems and there's going to be an overlap between Ai and web 3 as we venture into this new space after uh chat chat gbt has launched and what I think about responsible AI is that we as scientists developers Builders we have an intrinsic responsibility to ensure that we identify the encoded biases in the algor in the algorithms that we create and we also follow standards and assessments and create those standards and assessments so that we can have that access across the board and so here's a couple case studies um there's a man named dwan buck and he was put on death row he was put on death row because a prosecutor in Houston said that because he was black he was more likely to commit a crime and that was based off a survey called an lsir survey that leverages Ai and ML and some of those questions in that LS I our survey are like how many times have you been pulled over by the police how many times have you been convicted well that is a direct relationship with racial bias that exists in technology and so this is a extreme example of how weapons of math destruction can lead to someone losing their life then we'll also talk about the invasive side the Privacy side Target there was a a data scientist that sitting down at his desk and two marketers come up to him and say hey man can you make an algorithm for me to figure out if someone is pregnant or not based upon pieces of text cookies right the things they're looking on at the we on their uh Target website and he said sure I'll make this algorithm and funny enough a man one day goes into his home and he sees on the coffee table a magazine and a Target magazine and it's directed for his daughter and his daughter is underage but he's trying to figure out why is Target sending his daughter a magazine that has baby clothes that has cotton balls for children so he contacts Target and says why are you sending this for to my kid and that's how he found out that his child was pregnant and then also recently we've saw push back with this San Francisco has um they put out a bill to try to allow robots to work with police officers now of course they retracted that they said that they came to the conclusion that it would disproportionately affect black and brown people I mean of course but this is the direction that we're heading if we don't slow down when it comes to this Innovation with the security assessments and the standards and so one of my favorite scientists Dr timet gibu she uh was Ed from Google uh years ago and she spoke about the existential risks in Ai and people shut her out and they didn't really listen to what she had to say because of the contentiousness of her being ousted from Google but now she has an institution called D Institute where they're working FR uh very hard when it comes to the ethical Ai and the responsible AI um topic and she says that scientists are some of the most dangerous people in the world because we have this illusion of objectivity and so that brings me back to my case and my point as we Galvanize at e Denver and we built this Tech we also have to be cognizant of the encoded biases that we may or may not be propagating in our in our technology and in our creations and so here are some book recommendations and also a Netflix series that I really liked uh coded bias if you guys have time where you guys are coding tonight check out coded bias on Netflix it's created by Dr Joy who is the next woman on the um on for this next book unmasking AI she just launched that book and I think it's super excellent also weapons of math destruction by Kathy O'Neal Dr Kathy O'Neal she actually talks about the overlap a little bit of the overlap of why we even exist in this space as crypto um she talks about the Global Financial crisis and how we had leveraged uh these machine learning algorithms to decide who gets this loan who gets this house loan Etc and then we see the um the economy collapse Etc so there is kind of an overlap with the fluctuation of AI propagating and also the creation of our space which is the crypto blockchain space and I think that is like one of my most favorite books when it comes to ethical Ai and and also wired they have this magazine talking about timet uh timet gw's story and I definitely think that everyone should pick this up find a pdf version Etc but we are here in the future of how AI is integrated with a plethora of Technologies thanks to chat GPT and the hype between large language models thank you all very much my name is Jasmine Brunson I also go by bubbly Orca and I appreciate you all for listening to my talk remember don't get wrecked get rectified