All tracked items across vulnerabilities, news, research, incidents, and regulatory updates.
Keras versions up to 3.13.2 have a security flaw that lets attackers read files from your computer by hiding malicious instructions in model files. When you load a model using Keras functions like `keras.models.load_model()`, the vulnerability doesn't properly check if datasets are virtual (pointing to other files), so it automatically reads files an attacker specifies, potentially exposing sensitive data.
Fix: Fixed in versions 3.12.2 and 3.14.1.
NVD/CVE DatabaseNVIDIA Triton Inference Server for Linux has a use-after-free vulnerability (a bug where software tries to access memory that has already been freed, potentially causing crashes). An attacker could exploit this to cause a denial of service (making the service unavailable to legitimate users), though the severity rating from NIST has not yet been provided.
CVE-2026-24264 is a vulnerability in NVIDIA Triton Inference Server for Linux where an attacker can exploit improper handling of highly compressed data (data that has been reduced in size), potentially causing a denial of service (making a system unavailable to users). The vulnerability has a CVSS score of 4.0, which indicates a moderate severity rating.
Palantir CEO Alex Karp criticized the token model (a pricing system where AI companies charge based on the number of tokens, or text units, processed) used by OpenAI and Anthropic, saying costs have become unreasonably high and enterprises are losing interest. In response, businesses are shifting toward open weight models (AI systems with publicly available internal parameters that can be customized and run independently) and building their own custom AI tools to reduce expenses and maintain control over their data.
Cursor, an AI code editor used by over half of Fortune 500 companies, had two critical flaws (CVE-2026-50548 and CVE-2026-50549, both rated 9.8/10 severity) that allowed attackers to use prompt injection (hiding malicious instructions in data the AI reads) to escape the sandbox (a restricted environment limiting what commands can access) and run any command on a developer's computer without requiring any user action. The attacks worked by tricking the AI into writing to restricted system files, either by abusing a folder parameter or exploiting a flaw in how the editor checked for symbolic links (shortcuts that point to files).
A use-after-free vulnerability (UAF, where code accesses memory that has already been freed) exists in the Linux kernel's fhandle code because the may_decode_fh() function reads mount namespace data without holding locks, allowing another thread to free that memory concurrently. The bug only triggers if certain kernel configuration options are enabled and requires specific timing conditions.
Video game consoles like PlayStation 5, Xbox Series X/S, and Nintendo Switch 2 are becoming more expensive because AI datacentres are competing for the same computer chips and memory that consoles need. The demand for semiconductors (the tiny electronic components that power devices) has skyrocketed, especially after OpenAI made a deal to buy a huge portion of DRAM (the type of memory that stores data temporarily) from major manufacturers, causing prices to jump by as much as 200% and leaving console makers with fewer affordable options.
Google released a new smart speaker designed specifically for Gemini (Google's AI assistant), marking the company's first new smart speaker in six years. While the hardware is well-designed, the article indicates that Gemini for Home, the AI software powering the speaker, still feels incomplete and not fully ready for users.
Abu Dhabi's MGX, a major investment fund, has closed a $49 billion fund to back AI companies, making it one of the largest investment vehicles in the sector. The fund has invested in major AI companies like OpenAI and Anthropic, and is looking to invest across the AI tech stack (the layers of technology needed to build AI systems, including hardware, software, and platforms). This reflects the massive amount of money flowing into AI companies, which have raised a record $416.6 billion so far this year.
Fix: Karp suggests that open weight models and custom proprietary tools built by enterprises themselves, paired with partnerships like Palantir's expanded collaboration with Nvidia, offer a solution by giving companies 'control over their compute, their models, their data stack' and allowing them to 'own the means of production' rather than relying on expensive external AI labs.
CNBC TechnologyLLMs (large language models, AI systems trained on vast amounts of text) sometimes invent fake website domain names for real brands when answering questions, a problem called hallucination (generating false information that sounds plausible). Attackers can register these fake domains and use them for malicious activities, and this threat is hard to detect because the domains sound legitimate.
Fix: Both bugs are patched in Cursor 3.0, released April 2. All versions before 3.0 are affected, so users should update immediately.
The Hacker NewsLarge language models (AI systems trained on text to generate human-like responses) tend to give repetitive, predictable answers to open-ended questions, all converging on similar responses rather than showing creativity. A startup called Springboards built an LLM called Flint that was trained to produce a wider variety of responses to these types of questions, demonstrating that models can be made less predictable and more diverse in their outputs.
Fix: The startup Springboards has built an LLM called Flint, which has been trained to come up with a wider variety of responses than mainstream LLMs to open-ended questions. According to the source, Flint demonstrates greater diversity in outputs compared to models like ChatGPT and Claude.
MIT Technology ReviewFix: Fix it by taking rcu_read_lock() (a read-side lock for RCU, a synchronization mechanism that protects concurrent access) around the mount::mnt_ns access, like in __prepend_path(). Additionally, document the semantics of mount::mnt_ns, and use WRITE_ONCE() (a compiler directive that ensures a value is written exactly once without optimization) for writers that can race with lockless readers.
NVD/CVE DatabaseThis research examines security vulnerabilities in clustered distributed storage systems (CDSSs, which are networks of computers that store data across multiple machines organized into groups), focusing on attacks by two types of adversaries: omniscient adversaries (attackers with complete knowledge of the system) and limited-knowledge adversaries (attackers with partial information). The study analyzes how these attackers can compromise the system by sending corrupted data during file reconstruction and repair, and proposes theoretical bounds on system resilience using coding techniques.
Neural Radiance Fields (NeRF, a technique for representing 3D objects and scenes) need better ways to track which datasets were used to train them, especially since existing watermarking methods require changes during training that don't work well for this purpose. Researchers introduced TraceNeRF, a new method that embeds owner-specific identifying marks directly into training datasets using a hybrid approach combining spatial masking and discrete cosine transform (a mathematical technique for breaking images into frequency components), along with special modules to make these marks harder to remove or alter.
LiDAR subsystems (3D sensing systems that help autonomous vehicles perceive their surroundings) in self-driving cars face network-layer security risks when data packets travel through vehicle networks to the autonomous driving system. This study introduces a framework to trace how these network-level vulnerabilities propagate through the LiDAR data processing pipeline in Apollo V8.0 (a self-driving software platform), identifying three critical processing mechanisms (Frame Partition, Object Erasure, and Object Mark) that determine whether attacks succeed, and showing that failure depends on factors like timing and frame alignment rather than simple data corruption.
Researchers developed PromptFishing, a technique to identify whether social media accounts are controlled by LLMs (large language models, AI systems trained on text) or humans by embedding hidden prompts into normal conversations that cause LLMs to produce distinctive responses called hallucinations (when an AI generates false or off-topic information). The method uses a two-stage algorithm to create adversarial suffixes (specially crafted text additions) that trigger these distinctive LLM behaviors while appearing like normal conversation, achieving over 85% accuracy at detecting LLMs while falsely flagging humans less than 1% of the time.
This paper introduces SENSA (Self-Evolving Neuro-Symbolic Autonomy), a system that combines neural learning (data-driven pattern recognition) with symbolic reasoning (rule-based logic) to help AI make better decisions in complex, changing environments. The system uses causal reasoning (understanding why things happen, not just what happens) and can automatically improve its decision-making policies while staying robust against adversarial attacks (attempts to trick or disrupt the AI).
Researchers discovered a new ransomware tool called InfernoGrabber v9.0, created using the DeepSeek AI model, that exploits a legitimate browser feature (the File System Access API in Chrome and Chromium-based browsers) to encrypt files and demand ransom payments entirely within the browser on Windows and Android devices. This marks the first time an AI model has independently created a practical ransomware attack that bypasses browser sandboxing (the security feature that isolates browser processes), and it demonstrates that threat actors no longer need deep expertise to discover new attack methods.
Anthropic announced Claude Science, a new AI product designed to help scientific researchers conduct work autonomously in areas like computational biology and drug development, similar to how Claude Code assists software engineers. The US also lifted restrictions on Anthropic's Mythos and Fable models after security discussions, though the delay has already benefited Chinese AI competitors.
Researchers discovered that SemAder, a technique that manipulates both the structure and meaning of binary code (compiled programs), can fool LLM-based binary code analysis tools into missing security problems. The study shows that by carefully modifying how code is organized and what it semantically does, attackers can evade detection systems that use large language models to analyze compiled programs for vulnerabilities.
Check Point Research discovered a Python Flask web application where an AI model independently combined a browser vulnerability with a working ransomware technique to create malware that operates entirely within a web page, requiring no exploit kit, software installation, or technical skill from the attacker. This represents a shift in AI-assisted threats, as the AI connected theoretical security risks to practical attack methods on its own.