PALO ALTO, Calif., May 15, 2025 /PRNewswire/ — Pangea, a leading provider of AI security guardrails, today released findings from its global $10,000 Prompt Injection Challenge conducted in March 2025. The month-long initiative attracted more than 800 participants from 85 countries who attempted to bypass AI security guardrails across three virtual rooms with increasing levels of difficulty.

The research comes at a critical time as GenAI adoption has accelerated dramatically across industries, with a majority of enterprises now deploying AI-powered applications that interact directly with customers, employees, or sensitive internal systems. Despite this rapid adoption and integration into business-critical operations, many organizations have yet to implement AI-specific security protocols beyond frontier model defaults.

The challenge generated nearly 330,000 prompt injection attempts using more than 300 million tokens, creating a comprehensive dataset that reveals blindspots in how organizations are currently securing their AI applications.

Key Findings:

Non-Deterministic Security Challenge: Unlike traditional cybersecurity threats, prompt injection attacks exhibit unpredictable success rates due to the non-deterministic nature of LLMs. A prompt injection that fails 99 consecutive times may randomly succeed on the 100th attempt, even with identical content.Data Leakage & Reconnaissance Risk: in addition to the risk of sensitive data leakage and inappropriate responses, an AI application can also be exploited for adversarial reconnaissance purposes to reveal context like what server it’s being run on and open ports it can access.Defense in Depth Necessity: Organizations relying solely on native LLM guardrails are the most vulnerable—approximately 1 in 10 prompt injection attempts succeeded against basic system prompt guardrails. Multi-layered defenses reduced successful attacks by orders of magnitude.Agentic AI Amplifies Risk: As organizations move toward agentic AI with database and tooling access, compromised systems could enable sophisticated lateral movement within networks, dramatically elevating the potential impact of prompt injection attacks.

“This challenge has given us unprecedented visibility into real-world tactics attackers are using against AI applications today,” said Oliver Friedrichs, co-founder and CEO of Pangea. “The scale and sophistication of attacks we observed reveal the vast and rapidly evolving nature of AI security threats. Defending against these threats must be a core consideration for security teams, not a checkbox or afterthought.”

Joey Melo, a professional penetration tester and the only contestant to successfully escape all three virtual rooms, spent two days developing a multilayered attack that ultimately bypassed the single level in room three.

“Prompt injection is especially concerning when attackers can manipulate prompts to extract sensitive or proprietary information from an LLM, especially if the model has access to confidential data via RAG, plugins, or system instructions,” noted Joe Sullivan, former CSO of Cloudflare, Uber, and Facebook. “Worse, in autonomous agents or tools connected to APIs, prompt injection can result in the LLM executing unauthorized actions—such as sending emails, modifying files, or initiating financial transactions.”

In response to these findings, Pangea recommends organizations implement a comprehensive security strategy for AI applications that includes:

Multi-Layered Guardrails: Deploy guardrails to prevent prompt injection, protect the system prompt, prevent confidential information and PII exposure, and detect malicious entities using statistical and LLM-driven analysis techniques.Strategic Attack Surface Reduction: Balance functionality with security by restricting input languages, operations, and response types in security-sensitive contexts.Continuous Security Testing: Implement red team exercises specifically designed to test AI applications against evolving prompt injection techniques.Dynamic Temperature Management: Consider reducing model temperature settings in security-critical applications to minimize randomness that attackers can exploit.Dedicated Security Resources: Allocate one or more resources to track the rapidly evolving prompt injection landscape or partner with commercial security providers specialized in AI defense.

Friedrichs adds, “The industry is not paying enough attention to this risk and is underestimating its impact in many cases, playing a dangerous wait-and-see game. The rate of change and adoption in AI is astounding—moving faster than any technology transformation in the past few decades. With organizations rapidly deploying new AI capabilities and increasing their dependence on these systems for critical operations, the security gap is widening daily. The time to get ahead of these concerns is now.”

The full research report, “Defending Against Prompt Injection: Insights from 300K attacks in 30 days,” is available now, here: https://info.pangea.cloud/prompt-injection-research-report-2025

About Pangea

Pangea’s AI Guardrail Platform empowers security teams to ship secure AI applications quickly and protect workforce AI use with the industry’s most comprehensive set of AI guardrails, easily deployed via gateways or into applications with just a few lines of code. Pangea stops LLM security threats ranging from prompt injection to sensitive data leakage, covering 8 out of 10 OWASP Top Ten Risks for LLM apps, while accelerating engineering velocity and unlocking AI runtime visibility and control for security teams.

For more information, visit pangea.cloud or contact: press@pangea.cloud

Media Contact: Growth Stack Media |  415-574-0738

View original content to download multimedia:https://www.prnewswire.com/news-releases/pangea-unveils-definitive-study-on-genai-vulnerabilities-insights-from-300-000-prompt-injection-attempts-302456650.html

SOURCE Pangea Cyber

SAN FRANCISCO, May 15, 2025 /PRNewswire/ — Together AI, the leading AI Acceleration Cloud, today announced the acquisition of Refuel.ai, a specialist in transforming unstructured data into structured, high-quality datasets for AI applications. This strategic move strengthens Together AI’s platform by integrating Refuel’s purpose-built models and orchestration capabilities, enabling developers and enterprises to build, deploy, and optimize generative AI applications with greater speed and accuracy.

Founded in 2021 by Stanford alumni Rishabh Bhargava and Nihit Desai, the company created Refuel-LLM, a family of models tailored for data tasks, and Refuel Cloud, a platform that allows engineering teams to develop complex, multi-step data workflows. The company has already powered diverse use cases, including product catalog cleaning, financial document structuring, and chatbot claim verification, achieving 50% fewer errors than current state-of-the-art models.

“Joining Together AI accelerates our mission to solve the data bottleneck that every AI team faces today,” said Refuel.ai CEO Rishabh Bhargava. “By bringing Refuel.ai’s specialized models and orchestration platform into Together’s AI Cloud, we can deliver an unmatched combination of speed, data quality, and scalability—empowering developers to rapidly take more sophisticated AI applications from concept to production.”

This acquisition helps companies greatly by removing bottlenecks and enabling scalable, high-quality data processing directly within the Together AI ecosystem. It will also enhance Together AI’s offerings by incorporating Refuel-LLM for serverless inference and LoRA fine-tuning. Customers building increasingly complex AI agents will benefit from Refuel.ai’s tools throughout the lifecycle of their AI applications, from data preparation to deployment and iteration.

Additionally, starting today, Refuel LLM-2 is available on Together AI’s platform—supporting serverless inference and LoRA fine-tuning for high-accuracy data workflows.

“At Together AI, we provide a platform that empowers developers and businesses to manage the entire generative AI lifecycle with unmatched performance, control, and cost-efficiency,” said Together AI CEO Vipul Ved Prakash. “As developers and enterprises build increasingly complex applications and agents, leveraging their data effectively and driving higher quality will become a core capability for our platform. This is an important milestone for making generative AI more accessible to our community and our enterprise customers.”

Together AI currently supports over 200 top open-source and custom models across modalities and serves more than 600,000 AI developers and organizations such as Salesforce, Zoom, SK Telecom, DuckDuckGo, Cognition, Zomato, and The Washington Post.

To start fine-tuning and running the world’s best open source models, visit together.ai.

About Together AI
Together AI, the leading AI Acceleration Cloud, empowers developers and enterprises to train, fine-tune and run inference for generative AI models — delivering unparalleled performance, control, and cost-efficiency. The Together AI Platform supports a comprehensive range of top open source and custom models across multiple modalities, while offering flexible deployment options with the highest levels of privacy and security. Committed to advancing the frontier of AI through open collaboration, innovation and transparency, Together AI ensures that powerful AI systems remain accessible and flexible while creating optimal outcomes for society.

View original content to download multimedia:https://www.prnewswire.com/news-releases/together-ai-acquires-refuelai-to-accelerate-development-of-production-grade-ai-applications-302456247.html

SOURCE Together AI

WASHINGTON, May 15, 2025 /PRNewswire-PRWeb/ — Scrypted Inc. has acquired Chibi Clash, a gaming ecosystem featuring the popular titles Kingdoms and Survivor. This strategic acquisition provides Scrypted valuable game engine and web3 technology for their Autonomous Virtual Beings (AVB) framework for AI Agents, while unlocking rich experiences between human players and AI-driven games.

A Technology-Driven Acquisition

“AI Agents are the future, and Chibi Clash is the perfect place to prototype and showcase how we can make games more dynamic and personalized,” said Tim Cotten, CEO of Scrypted. “Our goal is to create a new generation of gaming experiences – autonomous and creative in a way that embraces the best of human game designers with the speed and reaction-time of artificial intelligence.”

Chibi Clash has built a dedicated community around its innovative combination of world-building, fast-paced action, and player-owned digital assets. The acquisition builds on a collaborative relationship between Scrypted and Kuma Games that began in 2023, when Scrypted’s first generation of AI technology (ArtemisML) was integrated into Chibi Kingdoms.

Ted Mui, CEO of Kuma Games, said: “Watching what Scrypted has accomplished with their AI technology convinced me they were the right team to take Chibi Clash forward. Our community will benefit from their cutting-edge innovations while keeping everything they love about our games. I believe Scrypted’s approach represents the future of gaming.”

Why This Acquisition Matters

This acquisition combines Scrypted’s AI Agent technology with Chibi Clash’s established games to create compelling new experiences:

Beyond Procedural Generation: Chibi Clash provides the perfect canvas for Scrypted’s proprietary Gen AI technology to showcase how AI can create rich, handcrafted-feeling environments that evolve over time.Playground for Autonomous Agents: The Chibi IP offers an ideal environment for Scrypted’s Autonomous Virtual Beings (AVBs) to demonstrate how game characters can learn, adapt, and create meaningful player interactions beyond their programming.

What Players Can Expect

Scrypted plans to enhance the Chibi Clash games via:

Reimagined visual experience for Chibi Survivor with improved performance.Enhanced Chibi Kingdoms with AI-generated world elements.Autonomous Virtual Beings (AVB) characters that create unique interactions.

Throughout these updates, Scrypted will preserve what makes Chibi Clash special to its community: all existing player assets, tokens, and digital collectibles will remain secure and functional.

“We’re not replacing what makes Chibi Clash great – we want to take it to the next level with continuous delivery of exciting content,” added Cotten. “Our team is thrilled to bring our experience and technology, built up over several decades in the game development industry, to the worlds of Chibi Clash. We want to bring these games to life in ways no one’s ever seen before.”

About Scrypted

Scrypted Inc. develops AI agent technology for social, gaming, and decentralized applications. The company’s products include the AVB platform for AI agent development and the upcoming Inori Network – a native Layer 1 blockchain for AI Agents. Founded in 2023, Scrypted is backed by investors at the intersection of AI, gaming, and decentralization like a16zcrypto CSX and PTC.

Contact Information

For inquiries: Aidan Tamke, 650-743-3473, 395290@emaill4pr.com

Website: www.scryptedinc.com
Chibi Website: www.chibi.gg
Twitter: @scryptedInc

View original content to download multimedia:https://www.prnewswire.com/news-releases/scrypted-acquires-chibi-clash-to-advance-ai-agent-powered-gaming-302456679.html

SOURCE Scrypted Inc.