<?xml version="1.0" encoding="utf-8" standalone="yes"?>
<rss version="2.0" xmlns:atom="http://www.w3.org/2005/Atom">
	<channel>
		<title>Prompt-Injection on dplabs — Software Engineering &amp; Technology Consultancy</title>
		<link>https://dplabs.tech/tags/prompt-injection/</link>
		<description>Recent content in Prompt-Injection on dplabs — Software Engineering &amp; Technology Consultancy</description>
		<generator>Hugo</generator>
		<language>en-us</language>
		
		
		
		
			<lastBuildDate>Tue, 21 Apr 2026 00:00:00 +0000</lastBuildDate>
		
			<atom:link href="https://dplabs.tech/tags/prompt-injection/index.xml" rel="self" type="application/rss+xml" />
			<item>
				<title>The Security Risks of AI-Enabled Applications</title>
				<link>https://dplabs.tech/blog/security-risks-ai-applications/</link>
				<pubDate>Tue, 21 Apr 2026 00:00:00 +0000</pubDate>
				<guid>https://dplabs.tech/blog/security-risks-ai-applications/</guid>
				<description>&lt;p&gt;Integrating AI into applications introduces security risks that traditional application security doesn&amp;rsquo;t address. Most security frameworks were designed for deterministic systems. LLMs are not deterministic — their outputs depend on inputs in ways that are difficult to predict, test, or constrain.&lt;/p&gt;&#xA;&lt;p&gt;This doesn&amp;rsquo;t mean AI applications are inherently insecure. It means you need to think about security differently when LLMs are in the loop.&lt;/p&gt;&#xA;&lt;h2 id=&#34;the-core-problem-llms-process-text-as-instructions&#34;&gt;The Core Problem: LLMs Process Text as Instructions&lt;/h2&gt;&#xA;&lt;p&gt;In a conventional application, there&amp;rsquo;s a clear distinction between code and data. A SQL injection attack tries to blur that boundary — to get user input interpreted as SQL commands. You prevent it by parameterizing queries, by treating user input as data rather than code.&lt;/p&gt;</description>
			</item>
	</channel>
</rss>
