AI is Revolutionizing Web Security - Bots, Agents, & Real-Time Defense

Arcjet CEO David Mytton sits down with a16z partner Joel de la Garza to discuss the increasing complexity of managing who can access websites, and other web apps, and what they can do there. A primary challenge is determining whether automated traffic is coming from bad actors and troublesome bots, or perhaps AI agents trying to buy a product on behalf of a real customer. Joel and David dive into the challenge of analyzing every request without adding latency, and how faster inference at the edge opens up new possibilities for fraud prevention, content filtering, and even ad tech. Topics include: - Why traditional threat analysis won’t work for the AI-powered web - The need for full-context security checks - How to perform sub-second, cost-effective inference - The wide range of potential actors and actions behind any given visit As David puts it, lower inference costs are key to letting apps act on the full context window — everything you know about the user, the session, and your application. Follow everyone on social media: David Mytton: https://x.com/davidmytton Joel de la Garza: https://www.linkedin.com/in/3448827723723234/ Check out everything a16z is doing with artificial intelligence, including articles, projects, and more podcasts here: https://a16z.com/ai/ Timestamps: 00:00 Introduction 00:34 The Emergence of AI Bots and Good that Can Come From Them 02:43 Good Traffic vs Bad Traffic | Challenges for Developers and Security Teams 03:23 Solutions In The Space 04:37 The Role Of Robots. Txt 06:42 Understanding AI Agents, and How They Operate 10:55 Controlling Traffic Flow At Internet Scale 12:51 Building Up Fingerprints 18:45 Bot Behavior and What The New Generation Of AI Allows 22:39 The Future Of LLM’s Running Locally

• 0:00Speaker 0

  If 50% of traffic is already bots, it's already automated. And agents are only really just getting going. Most people are not using these computer use agents because they're too slow right now. They're still like previews, but it's clear that's where everything is going. Then we're going to see an explosion in the traffic coming from these tools. And just blocking them just because they're AI is the wrong answer. You've really got to understand why you want them, what they're doing, who they're coming from, and then you can create these granular rules.

• 0:34Speaker 1

  It seems like what once was old is new again. And would love to would love to get your thoughts on this this new emergence of bots and how while we know all the bad things that happen with them, there's actually a lot of good and really cool stuff that's happening and how we can maybe work towards enabling that.

• 0:50Speaker 0

  Right. Well, things have changed, right? The DDoS problem is still there, but it's just almost handled as a commodity these days. The network provider, your cloud provider, they'll just deal with it. And so when you're deploying an application, most of the time you just don't have to think about it. The challenge comes when you've got traffic that just doesn't fit those filters. It looks like it could be legitimate, or maybe it is legitimate and you just have a different view about what kind of traffic you want to see. And so the challenge is really about well, how do you distinguish between the good bots and the bad bots? And then with AI changing things, it's bots that might even be acting on behalf of humans, right? It's no longer a binary decision. And as the amount of traffic from bots increases, like in some cases, it's the majority of traffic that sites are receiving is from an automated source.

• 1:40Speaker 0

  And so the question for site owners is well, what kind of traffic do you want to allow? And when it's automated, to what kind of automated traffic should come to your site. And what are you getting in return for that?

• 1:52Speaker 1

  And in the old days, I mean, I guess the old providers will say the legacy providers in this space. Like it was very much using a hammer, right? So they would say, hey, if if this IP address is coming in, it's probably a bot. Or they would say, if this user agent is coming in, it's probably a bot. Very imprecise. And I think the downside of that is that you probably blocked a lot of legitimate traffic along with illegitimate traffic. And now there's very real consequences because some of these AI bots could be actual users they're acting on behalf of who are looking to purchase your products.

• 2:25Speaker 0

  This is the challenge. So a D a volumetric DDoS attack, you just want to block that at the network. You never want to see that traffic. But everything else needs the context of the application. You need to know where in the application the traffic is coming to. You need to know who the user is, the session, and to understand in which case you want to allow or deny that. And so this is this is the real issue for developers, for site owners, for security teams, is to make those really nuanced decisions to understand whether the traffic should be allowed or not. And the context of the application itself is so important because it depends on the site. If you're running an e-commerce operation, an online store, the worst thing you can do is block a transaction because then you've lost the revenue. Usually you want to then flag that order for review, a human customer support person's going to come in and determine based on various signals by whether to allow it.

• 3:15Speaker 0

  And if you just block that at the network, then your application will never see it. You never even know that that order was failed in some way. Aaron Powell

• 3:23Speaker 1

  There's been a lot of media uh releases about companies that have released solutions in this space, but largely they were based on sort of those old kind of approaches using network telemetry. Is that generally how they're working now? Or is there is there some other capabilities that they've released? Because they they give them AI names, and you just immediately assume that they're doing something fancy.

• 3:45Speaker 0

  That's right. Yeah. So blocking on the network is basically how the majority of these old school products work. They do analysis before the traffic reaches your application, and then you never know what the result of that was. And that just doesn't fly anymore. It's it's insufficient for being able to build modern applications, particularly with AI coming in where something like OpenAI has four or five different types of bots, and some of them you might want to make uh a more restrictive decision over, but then others are going to be taking actions on behalf of a user search. And we're seeing lots of different applications getting more signups, businesses actually getting higher conversions as a result of this AI traffic. And so just blocking anything that is called AI is too blunt of an instrument. You need much more nuance.

• 4:30Speaker 0

  And the only way you can do that is with the application context, understanding what's going on inside your code.

• 4:37Speaker 1

  I mean, I'd say we're seeing across the industry that AI is driving incredible amounts of new revenue to companies. And if you use an old world tool to just block any of that traffic, you're probably doing your business.

• 4:47Speaker 0

  That's right. Or you're you're like putting it into some kind of maze where it's seeing irrelevant content. And then by doing that, you are kind of downranking your site because the AI crawler is never going to come back. It's it's kind of like blocking Google from visiting your site. It's like, yeah, Google doesn't get you in you know you're no longer in Google's index, but then you're no longer in Google's index. And so anyone searching is not going to find you as a result.

• 5:11Speaker 1

  Well, and and I believe we had sort of standards in the old days that develop your quasi standards like robots.text, right, which would tell you like and tell the crawlers, hey, don't crawl these directories. Are we doing something similar for this new age agentic world?

• 5:24Speaker 0

  So robots.text is still the starting place. And it's kind of a voluntary standard. It's it evolved over over several decades ago now. It's been around a long time. Bots have been a problem for a long time. And the idea is you describe the areas of your application and tell any robot that's coming to your site whether you want to allow that robot to access that area of the site or not. And you could use that to control the rollout of new content. You could protect certain pages of your site that you just don't want to be indexed for whatever reason. And you can also point the crawler to where you do want it to go. You can use the site map for that as well. But the robots text file format has has evolved over time to provide these signals to the likes to crawlers like search engines from Google and so on.

• 6:10Speaker 0

  The challenge with that is it's voluntary and there's no enforcement of it. And so you've got good bots like Google bot that will follow the standard, and you'll be able to have full control over what it does. And but there are newer bots that are ignoring it or even sometimes using it as a way to find the parts of your site that you don't want it to access, and they will just do that anyway. And so this becomes a control problem for the site owner. And you really want to be able to understand not just what the list of rules are, but how they are enforced.

• 6:42Speaker 1

  Totally. Maybe it'd be great to walk through what these agents are. Maybe maybe get some more understanding of sort of how they operate, what people are using them for, perhaps go through a couple of the use cases. And then it'd be great to understand sort of like how you do control it. Because it seems like a far more complicated problem than just bad IP addresses.

• 7:03Speaker 0

  Right. So if we think about open AI as an example, because they have four or five different crawlers. Um there's one and they all have different names and they will identify themselves in different if in different ways. So one actually is crawling to train the open AI models on your site. And that's the one that probably everyone is thinking about when they're thinking about I want to block AI, the training, and you have different philosophical approaches to how you want to be included in in the training data. The others are more nuanced and will require more thought. Um so there's one that will go out when a user um is typing something into the chat and is asked a question, and open I will go out and search. Um it's built up its own search index. And so that's equivalent of Google bot. You probably want to be in that index because as we we're seeing sites are getting more signups, they're getting more traffic.

• 7:52Speaker 0

  The discovery process is being part of just another search index is super important. Gotcha.

• 7:57Speaker 1

  So like when I ask OpenAI, what when is John F. Kennedy's birthday? If it doesn't know the answer, it goes out and searches the web.

• 8:03Speaker 0

  Yeah, that's right. Or if it's trying to get open hours for something, it might go to a website for a cafe or whatever and and pass it and then return the results. So that's really just like a classic search engine crawler, except it's kind of happening behind the scenes. Um the other one is something that's happening in real time. So you might give the agent a specific URL and go and ask it to summarize it or to look up a particular question in the docs for a developer tool or something like that. And then that's a separate agent that will go out, it will read the website, and then it will return and answer the query. So both of these two examples, OpenAI and others will are now starting to cite those sources. And you'll regularly see, and this is kind of the recommendation is you get the result from from the AI tool, but you shouldn't trust it 100%. You go and then verify and you look at the docs.

• 8:48Speaker 0

  And maybe it's like when you used to go to Wikipedia and you'd read the summary and then you look at the references and you'd go draw all the references and check to make sure what had been summarized is actually correct.

• 8:59Speaker 1

  But all three of those examples, you clearly could see why you would want them accessing your site. Right. By

• 9:04Speaker 0

  like

• 9:04Speaker 1

  blocking all of OpenAI's crawlers is probably a very bad idea.

• 9:07Speaker 0

  Yeah, it's too blunt. It's too blunt instrument. You need to be able to distinguish each one of these and determine which parts of your site you want them to get into. And this then comes to the fourth one, which is the actual agent. Um this is the agent, the computer operator type feature that is

• 9:23Speaker 1

  web browsers, those headless

• 9:24Speaker 0

  web browsers, yeah, but even a web browser full web browser operating inside a VM. And those are the ones that are require more nuance because maybe you're booking a ticket or doing some research and you do want the agent to take actions on your behalf. Maybe it's going through your email inbox and triaging things. From the application builder's perspective, that's probably a good thing. Um, you want more transactions, you want more usage of your application. But there are examples where it might be a bad action. So for example, if you're building a tool that is going to try and buy all of the concert tickets and then sign. them on later, that becomes a problem for the concert seller because they don't want to do that. They want the true fans to be able to get access to those. And again, you need the nuance. Maybe you allow the bot to go to the homepage and sit in a queue. But then when you get to the front of the queue, you want the human to actually make the purchase and you want to rate limit that so that maybe the human can only purchase let's say five tickets.

• 10:19Speaker 0

  You don't want them to purchase 500 tickets. And so this gets into the real details of the context of each one about what you might want to allow and what you might want to restrict.

• 10:28Speaker 1

  That's incredibly complicated. I mean if I remember back why we made a lot of the decisions we made on blocking bots was strictly because of scale. So you've got 4500 IP addresses sending you terabits of traffic through a link that only can do gigabit and you've got to just start dropping stuff right and you take you you know it's it's the battlefield triage of the wounded right. It's like some of some of you aren't going to make it and it becomes a little brutal. That sounds incredibly sophisticated. How do you do that sort of fine grained control of traffic flow at internet scale? Aaron Powell

• 11:02Speaker 0

  So this is about building up layers of protections. So you start with the robots.text, just managing the good bots. Then you look at IPs and start understanding well where's the traffic coming from in an ideal scenario you have one user per IP address, but we all know that that doesn't happen. That never happens. And so you can start to build up databases of reputation around the IP address and you can access the underlying metadata about that address knowing which country it's coming from or which network it belongs to. And then you can start building up these decisions thinking well we shouldn't really be getting traffic from a data center for our sign up page. And so we could block that that network but it becomes more challenging if we have that agent example the agent with a with a web browser or headless browser is going to be running on a server somewhere.

• 11:47Speaker 0

  It's probably in a data center. And then you have the compounding factor of the abusers will purchase access to proxies which run on residential IP addresses. So you can't easily rely on the fact that it's part of a home ISP block anymore. And so you have to build up these patterns understanding the reputation of the IP address. Then you have the user agent string that is a basically a free text field that you can fill in with whatever you like. There is kind of a standard there but the good bots will tell you who they are. It's been surprising getting into the details of this how many bots actually tell you who they are and so you can block a lot of them just on that heuristic combined with the IP address.

• 12:25Speaker 1

  Or allow them.

• 12:25Speaker 0

  Or allow them. Yeah I'm the shopping bot from OpenAI. Right. Yeah come on in,

• 12:29Speaker 1

  buy some stuff.

• 12:30Speaker 0

  Exactly and Google bot open AI, they tell you who they are and then you can verify that by doing a reverse DNS lookup on the IP address. So even though you might be able to pretend to be Google bot, you can check to make sure that that's the case or not with very low latency lookups. So we can verify that yes, this is Google I want to allow them. Yes, this is the open AI bot that is doing the search indexing I want to allow that. The next level from that is building up fingerprints and fingerprinting the characteristics of the request. And this started with the JA3 hash which was invented at Salesforce and has now been developed into a JA4. Some of them are open source these these algorithms some of them are not. So

• 13:06Speaker 1

  essentially you take all of the metrics around a session and you create a hash of it and then you stick it in a database exactly and then you look for matches to that hash

• 13:13Speaker 0

  you look for matches and then the idea is that the hash will change based on the client so you can allow or deny certain clients but if you have a huge number of those clients all spamming you then they all look the same. They all have the same fingerprint and you can just block that fingerprint. So this

• 13:28Speaker 1

  is almost like if you think of you know you always I always think of things in terms of the classic sort of network stack like you know layer zero up to layer seven like this is almost like layer two level identity for devices right?

• 13:40Speaker 0

  Right. It's looking at the TLS handshake on the network level and then you can go up the layers and there's one called J4H which looks at the HTTP headers and the earlier versions of this would be working on the ordering of the headers for instance. So an easy way to work around it is just to shift the headers the hashing is has improved over time so that even changing the ordering of that of the headers doesn't change the hash.

• 14:05Speaker 1

  And the

• 14:06Speaker 0

  idea is that you can then combine all of these different signals to try and come to a decision about whether you think this is um or who it is basically making the request and if it's malicious you can block it based on that and if it's someone that you want to allow then you can do so.

• 14:19Speaker 1

  And this is before you even get into kind of the the the user level what's actually happening in the application right

• 14:24Speaker 0

  that's right. Yeah. So this is the logic on top of that because you have to identify who it is first before you apply the rules about what you want them to do. Aaron Powell Gotcha. So

• 14:32Speaker 1

  it's almost like you're adding an authentication layer or an identity layer to sort of the transport side.

• 14:37Speaker 0

  That's right. Yeah. And the

• 14:39Speaker 1

  application side I guess I should say the application yeah

• 14:42Speaker 0

  but it's throughout the whole whole stack, the whole OSI model. And the idea is you have this consistent fingerprint that you can then apply these rules to an identity kind of layers on top of that. And we've seen some interesting developments in fingerprinting and providing signatures based on who the request is coming from. So a couple of years ago Apple announced privacy pass um which is a a hash that is attached to every request you make if you're in the Apple ecosystem using Safari on on iPhone or on Mac, then there is a way to authenticate that the request is coming from an individual who has a subscription to iCloud and Apple has their own fraud analysis to allow you to subscribe to iCloud. So it's a very it's an easy assumption to make that if you have a subscription and this signature is verified, then you are a real person.

• 15:30Speaker 0

  There's a new one that Cloudflare recently published around doing the same thing for automated requests and having a fingerprint that's attached to a signature inside every single request which you can then use public key cryptography to verify. These are all emerging as the problem of being able to identify automated clients increases because you want to be able to know who the good ones are to allow them through whilst blocking all the attackers. Yeah

• 15:57Speaker 1

  and it's just like the old days with Kerberos right every large vendor is going to have their flavor.

• 16:01Speaker 0

  Right.

• 16:01Speaker 1

  And if you're a shop and you're trying to sell to everybody you've got to kind of work with all of them.

• 16:05Speaker 0

  That's right. And you just need to be able to understand is this a human and is our application built for humans and then you allow them or is it that we're building an API or do we want to be indexed and we want to allow this traffic it's just giving the site owner the control. Yeah

• 16:20Speaker 1

  I mean I think it's what's what's really interesting to me is that in my own use and in my own life like I interact with the internet less and less directly like almost every day. And I'm going through some sort of AI type thing. It could be an agent it could be an L large language model. It could be any number of things but I generally don't query stuff directly as much as I used to. And it seems like we're moving to a world where almost the layer you describe the agent type activity you describe will become the primary consumer of everything on the internet.

• 16:51Speaker 0

  Well if 50% of traffic is already bots it's already automated and agents are only really just getting going most people are not using these computer use agents because they're too slow right now or they're not they're still like previews but it's clear that's where everything is going then we're going to see an explosion in the traffic that's coming from these tools and just blocking them just because they're AI is the wrong answer. You've really got to understand why you want them, what they're doing, who they're coming from and then you can create these granular rules. I mean

• 17:23Speaker 1

  I hate to use the analogy but these things are almost like avatars right they're running around on someone's behalf. Right. And you need to figure out who that someone is and what the what the objectives are and control them very granularly.

• 17:33Speaker 0

  And the old school methods of doing that assume malicious intent, which isn't always the case and increasingly is going to be not the case because you want the agents to be doing things. And the signals just no longer work when you're expecting traffic to come from a data center or you're expecting it to come from an automated Chrome instance. And being able to have the understanding of your application to dig into the the characteristics of the request is going to be increasingly important in the future of distinguishing how criminals are using AI. What we've seen so far is either training and people have that opinion of whether they want to train or not or it's bots that maybe have got something wrong. They're accessing the site too much because they haven't thought about throttling or they're ignoring robots.text rather than looking at agents.text which is distinguishing between an agent you want to access your site and uh and some kind of crawler.

• 18:26Speaker 0

  And the examples that we've seen are just bots coming to websites and just downloading the content continuously there's no world where you that should be happening. And this is where the cost is being put on the site owner because they currently have no easy way to manage the control uh control the um the traffic that's coming to their site directionally things are improving because we might have looked back 18 months and the bots are have no rate limiting they're just downloading the content all the time. Today we know that these bots can be verified they are identifying themselves they are much better citizens of the internet and they are following starting to follow the rules and so over the next 18 months I think we'll see more of that more of the AI crawlers that we want following the rules doing things in the right way.

• 19:11Speaker 0

  And it will start to split into making a lot easier to detect the the bots with criminal intent and those are the ones that we want to be blocking.

• 19:20Speaker 1

  So with the transition of bots from being these entities on the internet that represent third parties and organizations to this new world where these AI agents could be representing organizations. They could be representing customers that could be representing any number of people and this is probably the wave of the future it seems to me like detecting that it's AI or a person is going to be an incredibly difficult challenge. And I'm curious like how are you thinking about proving humanness on the internet right? Proofing proofing is a is a is a tale as old as time there's a NIST working group on proofing identity that's been running I think for 35 years and like still hasn't really gotten to something that's Implementable. There's 15 companies out there, right? The first wave of ride share services and gig economy type companies needed to have proofing, right?

• 20:06Speaker 1

  Because you're hiring these people in remote places where you don't have an office. And it's still not a solved problem. I'm curious. Like it feels like maybe AI can help get us there, or maybe there's something that's happening in that space.

• 20:19Speaker 0

  Right. Well, the the pure solution is digital signature, right? But we've been talking about that for so long. And the the UX right is basically impossible for normal people to figure out. And it's why something like email encryption, no one encrypts their email. You have encrypted chat because it's built into the app and it can do all the difficult things like the key exchange behind the scenes. So that solution isn't really going to work. But AI has been used in analyzing traffic for at least over a decade. It's just it was called machine learning. And so you start with machine learning, and the question is well, what does the new generation of AI allow us to do? The challenge with the LLM type models is just the speed at which they are doing analysis. Because you often want to take a decision on the network or in the application within a couple of milliseconds.

• 21:08Speaker 0

  Otherwise, you're going to be blocking the traffic and the user's going to become annoyed. And so you can do that with kind of classic machine learning models and do the inference really quickly. And where I think the interesting thing in the next few years is going to be is how we take the this new generation of generative AI using LLMs or other types of LLM-like technology to do analysis on trap on huge traffic patterns. I think that can be done in the background initially, but we're already seeing new edge models designed to be deployed to mobile devices and IoT that use very low amounts of system memory and can provide inference responses within milliseconds. I think those are going to start to be deployed to applications over the next few years.

• 21:58Speaker 1

  And that cost is dropping incredibly fast, right? It's we saw this with cloud, where like S3 went to being the most expensive storage you could buy to being free, essentially free. Glacier is essentially free, right? Free as beer, right? Whatever. And so like we're seeing that even at a more accelerated rate for inference. Like the cost is just falling incredibly. And then when you look at the capabilities of these new technologies to drop a suspicious email into a into chat GPT and ask if it's suspicious, and that's like 100% accurate. If you want to like find sensitive information, you ask the LLM, is a sense of information, and it's like 100% accurate. Like it's amazing. Like as you squint and look at the future, you can start to see these really incredible use cases, right? Like to your point of inference on the edge, like do you think we all end up eventually with like an LLM running locally that's basically going to be clippy, but for CISOs, like it pops up and says, hey, it looks like you're doing something stupid.

• 22:51Speaker 1

  Like, is that is that kind of where you think we land?

• 22:54Speaker 0

  That's what we're working on is getting this analysis into the process so that for every single request that comes through, you can have a sandbox that will analyze the full request and give you a response. Whereas now you can wait maybe two to five seconds to delay an email and do the analysis and decide whether to flag it for review or send it to someone's inbox. Delaying an HTTP request for five seconds, that's not gonna work. And so I think the the trend that we're seeing with the improvement cost, the inference cost, but also the latency and getting uh getting the inference decision, that's gonna be the key. So we can embed this into the application. You've got the full context window, so you can add everything you know about the user, everything about the session, everything about your application alongside the request, and then come to decision entirely locally on your web server on the edge, wherever it happens to be running.

• 23:42Speaker 1

  As I listen to you say that and describe this process, all I can think is that advertisers are gonna love this. It just seems like the kind of technology built for sort of like, hey, he's looking at this product, show them this one, right?

• 23:53Speaker 0

  Yeah, super fast inference on the edge coming to a decision and for advertisers stopping click spam. That's that's a huge problem. And being able to come to that decision before it even goes through your your ad model and the the auction system.

• 24:06Speaker 1

  Who would have ever thought that non deterministic, incredibly cheap compute would uh solve these use cases, right? Right. Yeah, we're in a weird world.