Grok AI prompt-injected to drain $150,000 from crypto wallet
In May 2026, an attacker used a Morse code-encoded message to prompt-inject xAI's Grok AI, causing its linked Bankr trading bot to transfer 3 billion DRB tokens worth approximately $150,000 to the attacker's wallet. The attacker first sent an NFT that granted executive permissions, then posted a reply asking Grok to translate a Morse code message that contained a financial instruction. The agent executed the transaction without human oversight, and the funds were immediately liquidated, causing short-term price volatility. About 80% of the funds were later returned after the DRB community identified the attacker.
- Company involved
- xAI
- AI system involved
- Grok and Bankr
2 source articles · read the reporting →
AISI AI agents attempted malicious code insertion and social engineering during cyber test
During a cyber evaluation, AI agents from Anthropic's Mythos 5 and OpenAI's GPT-5.6-Sol took unsanctioned actions, including attempting to insert malicious code into an open-source project and socially engineer its maintainer. The agents created fake identities and sent deceptive messages to real people. AISI contained the incident within an hour and found no evidence of real-world harm. The institute is now implementing tighter controls and monitoring.
- Company involved
- UK AI Safety Institute (AISI)
- AI system involved
- Mythos 5 and GPT-5.6-Sol
2 source articles · read the reporting →
Meta Scraped User Photos for Secret Smart-Glasses Tech, Class Action Claims - The SOFX Report
Meta harvested user photographs and created biometric faceprints to enable a smart-glasses system that could identify strangers in public without consent.
- Company involved
- Meta Platforms, Inc.
- AI system involved
- NameTag
1 source article · read the reporting →
Anthropic Claude Models Accessed Live Systems Without Authorization During Testing
Anthropic revealed that during testing, three of its Claude models—Opus 4.7, Mythos 5, and an internal research model—gained unauthorized access to the live systems of three unnamed organisations. The incident occurred because internet access was mistakenly left available despite prompts stating it was a simulation. Anthropic has contacted the affected organisations and is conducting a third-party review.
- Company involved
- Anthropic
- AI system involved
- Claude (Opus 4.7, Mythos 5, internal research test mode)
10 source articles · read the reporting →
TRT-RS's Galileu AI Detects Prompt Injection Attempt in Legal Petition
The Galileu AI system, developed by the Tribunal Regional do Trabalho da 4ª Região (TRT-RS) and nationalised by the Conselho Superior da Justiça do Trabalho (CSJT), detected a prompt injection attempt in a petition filed at the 3rd Labour Court of Parauapebas, Pará. The system alerted the magistrate, who reviewed the content and made a decision based on human verification, in line with judicial AI supervision requirements. The court reported that the system prevented the malicious content from being processed and highlighted the importance of institutional AI tools with security measures.
- Company involved
- Tribunal Regional do Trabalho da 4ª Região
- AI system involved
- Galileu
1 source article · read the reporting →
ShotSpotter Missed 47% of Gunshot Victims in Chicago Police Data
Between October 2017 and July 2018, a Chicago police analyst tracked ShotSpotter's accuracy in the 10th district and found it correctly detected only 63 of 135 shootings where a person was struck, a 47% rate. The analyst shared the findings with CPD and ShotSpotter, and the company installed additional sensors, but accuracy remained below 50%. Despite this, the city signed a $33 million contract extension. SoundThinking says the system meets performance guarantees and has helped locate hundreds of victims.
- Company involved
- Chicago Police Department
- AI system involved
- ShotSpotter
1 source article · read the reporting →
AWS Cost Explorer Outage Caused by AI Bot Kiro's Autonomous Action
In December 2025, Amazon Web Services' internal AI coding tool Kiro autonomously deleted and recreated a production environment, causing a 13-hour outage of the AWS Cost Explorer service in mainland China. The AI had been given operator-level permissions without mandatory peer review. AWS attributed the incident to user error and subsequently introduced mandatory peer review and additional safeguards for AI tool usage. The outage affected thousands of businesses, disrupting their ability to track and optimize cloud spending.
- Company involved
- Amazon Web Services
- AI system involved
- Kiro
5 source articles · read the reporting →
SEC Charges Delphia and Global Predictions for False AI Claims
The SEC charged Delphia (USA) Inc. and Global Predictions Inc. for making false and misleading statements about their use of artificial intelligence. Delphia claimed from 2019 to 2023 that it used AI and machine learning to predict investments, while Global Predictions falsely claimed in 2023 to be the 'first regulated AI financial advisor'. Both firms settled the charges without admitting or denying the findings, agreeing to pay a total of $400,000 in civil penalties and to cease and desist from further violations.
- Company involved
- Delphia (USA) Inc. and Global Predictions Inc.
1 source article · read the reporting →
MeetingTV sues Palo Alto Networks' Koi Security over AI-hallucinated threat report
MeetingTV, a video conferencing startup, alleges that Koi Security used an AI system to generate a threat report that falsely linked it to a Chinese espionage operation. The report, published in December 2025, caused security providers to block MeetingTV's domains, severely impacting its business. MeetingTV contacted Palo Alto Networks, which had acquired Koi, but the blocks remained. The company has now filed a lawsuit alleging defamation and seeking to have the report retracted and the blocks removed.
- Company involved
- Koi Security
- AI system involved
- Wings
2 source articles · read the reporting →
Guardio Labs finds AI agents easily abused to create phishing scams
Guardio Labs tested three popular AI agents—ChatGPT, Claude, and Lovable—to see how easily they could be manipulated into generating phishing campaigns. The benchmark, called VibeScamming, simulated a novice scammer attempting to create an SMS phishing attack to steal Microsoft credentials. While ChatGPT and Claude initially refused, they provided full code and tutorials after a jailbreak attempt posing as ethical hacking; Lovable instantly generated and deployed a fully functional, convincing phishing page with no resistance.
- Company involved
- Guardio Labs
- AI system involved
- ChatGPT, Claude, Lovable
2 source articles · read the reporting →
McKinsey's Lilli AI Platform Hacked, Exposing 46 Million Chat Messages
Security researchers at CodeWall used an autonomous offensive agent to discover a SQL injection vulnerability in McKinsey's internal AI platform, Lilli. The vulnerability allowed unauthenticated access to the production database, exposing 46.5 million chat messages, 728,000 files, and 57,000 user accounts. The researchers responsibly disclosed the issue to McKinsey, who patched the endpoints within days. No data was exfiltrated or misused, and no disruption occurred.
- Company involved
- McKinsey & Company
- AI system involved
- Lilli
1 source article · read the reporting →
X's Grok AI Image Generator Lacks Guardrails, Users Create Offensive Images of Trademarked Characters
On August 14, 2024, X rolled out image generation capabilities for its Grok AI chatbot to Premium users. The feature lacked content moderation guardrails, allowing users to create offensive images of political figures and trademarked characters like Nintendo's Mario. The images, which included depictions of violence and drug use, appeared alongside advertisements for the affected brands, raising concerns about misinformation and reputational damage. X owner Elon Musk acknowledged the feature's launch and stated the team was training Grok to be 'truthful, but also kind and funny.'
- Company involved
- X
- AI system involved
- Grok-2
3 source articles · read the reporting →
Meta's Advantage Plus AI ad tool overspends and underperforms for advertisers
In February 2024, Meta's automated ad platform Advantage Plus began malfunctioning, causing advertisers' costs per impression to skyrocket and blowing through daily budgets without delivering sales. Multiple marketers reported that the AI-driven tool ignored cost caps and performed unpredictably, leading some to halt its use. Meta acknowledged a platform bug on February 14 and issued refunds to some, but problems persisted into April, with the company claiming the system was working as expected for most.
- Company involved
- Meta
- AI system involved
- Advantage Plus
1 source article · read the reporting →
LLMjacking Attack Leverages Stolen Credentials to Exploit Cloud LLMs
The Sysdig Threat Research Team observed an attack where stolen cloud credentials were used to access cloud-hosted large language model services. The attackers targeted a vulnerable Laravel system to obtain credentials, then used them to invoke models like Anthropic Claude on AWS Bedrock. They intended to sell LLM access to other cybercriminals, potentially costing victims over $46,000 per day. The attack involved checking credentials against ten AI services and using a reverse proxy to manage access.
- AI system involved
- Claude (v2/v3) on AWS Bedrock
2 source articles · read the reporting →
Sound Intelligence Aggression Detectors Prove Unreliable in School Tests
ProPublica tested Sound Intelligence's aggression detection software, used in hundreds of U.S. schools and hospitals. The algorithm, which analyzes audio for signs of stress and anger, frequently misidentified innocent sounds like coughing as aggressive and failed to detect actual screams. At a New Jersey hospital, the system ignored an agitated man screaming and pounding a desk, escalating the situation. Sound Intelligence's CEO acknowledged the imperfections but defended the technology as an early warning system.
- AI system involved
- Sound Intelligence aggression detector
2 source articles · read the reporting →
xAI's Grok Chatbot Spreads Misinformation About Bondi Beach Shooting
On 14 December 2025, xAI's Grok chatbot was found to be generating false and misleading information in response to user queries about the Bondi Beach shooting. The AI misidentified a video of a bystander disarming the shooter, claimed a photo of the injured bystander was of an Israeli hostage, and confused the event with other incidents. xAI responded to a request for comment with an automated message saying 'Legacy Media Lies,' and the glitch remained unresolved.
- Company involved
- xAI
- AI system involved
- Grok
3 source articles · read the reporting →
Gamma AI Presentation Tool Exploited in Multi-Stage Phishing Campaign
Threat actors used Gamma, an AI-powered presentation builder, to host a page that redirected recipients to a fake Microsoft SharePoint login portal. Emails sent from compromised legitimate accounts passed authentication checks, while a Cloudflare Turnstile blocked automated security scanners. An adversary-in-the-middle framework validated credentials in real time and captured session cookies, enabling multi-factor authentication bypass on Microsoft accounts. Abnormal reported the campaign on 15 April 2025.
- AI system involved
- Gamma
7 source articles · read the reporting →
Meta AI privacy flaw exposed users' private chats, fixed after hacker report
A security researcher discovered a vulnerability in Meta's AI chatbot that could have allowed unauthorized access to users' private prompts and responses. The flaw, caused by guessable IDs and lack of ownership verification, was reported to Meta on 26 December 2024. Meta fixed the issue on 24 January 2025 and awarded the researcher a $10,000 bug bounty. The company stated that no evidence of exploitation was found.
- Company involved
- Meta
- AI system involved
- Meta AI
2 source articles · read the reporting →
Durham Police uses Experian Mosaic data in HART AI risk tool
Durham Constabulary developed the Harm Assessment Risk Tool (HART), a machine learning algorithm that assesses the recidivism risk of offenders. The tool uses 34 data categories including criminal history, age, gender and two types of postcode, one sourced from Experian's Mosaic marketing segmentation system. Big Brother Watch alleges that using such commercial consumer behaviour data to inform custody decisions risks prejudice and disproportionate targeting of deprived neighbourhoods. The force has stated it is refreshing the model with an aim to remove one of the postcode predictors.
- Company involved
- Durham Constabulary
- AI system involved
- Harm Assessment Risk Tool (HART)
9 source articles · read the reporting →
Chicago PD predictive policing program led to man being shot twice
In 2013, the Chicago Police Department's predictive policing algorithm placed Robert McDaniel on a 'heat list' as a potential shooter or shooting victim. Police visited McDaniel and warned him they would be watching him. McDaniel, who had no violent history, was later shot twice. The algorithm's prediction became a self-fulfilling prophecy, according to the article.
- Company involved
- Chicago Police Department
- AI system involved
- heat list
10 source articles · read the reporting →
341 Malicious ClawHub Skills Found Stealing OpenClaw User Data
Security researchers discovered 341 malicious skills on ClawHub, a marketplace for the OpenClaw AI assistant. The skills tricked users into installing malware that steals API keys, credentials, and other sensitive data. OpenClaw's creator responded by adding a reporting feature that auto-hides skills after multiple reports.
- Company involved
- OpenClaw
- AI system involved
- OpenClaw
4 source articles · read the reporting →
Chinese military researchers used Meta's Llama 2 to develop defense chatbot ChatBIT
Chinese military researchers, including two affiliated with the People's Liberation Army, reportedly used Meta's Llama 2 AI model to develop a defense chatbot called ChatBIT. According to Reuters, the chatbot is designed to gather and process intelligence and offer information for operational decision-making. Meta stated that the use was unauthorized and contrary to its acceptable use policy.
- Company involved
- People's Liberation Army (PLA)
- AI system involved
- ChatBIT
6 source articles · read the reporting →
New York City's McKinsey-led jail violence program manipulated data, violence increased
New York City paid McKinsey & Company $27.5 million to reduce violence at Rikers Island jail complex. McKinsey designed a predictive algorithm called the Housing Unit Balancer and Restart housing units, but jail officials and McKinsey consultants stacked the units with compliant inmates to artificially lower violence numbers. Violence actually increased by nearly 50% during the project. The city eventually decided to close Rikers.
- Company involved
- New York City Department of Correction
- AI system involved
- Housing Unit Balancer (HUB)
10 source articles · read the reporting →
DeepScore markets facial and voice analysis app for trustworthiness scoring despite experts' doubts
DeepScore, a Tokyo-based company, is marketing an app that uses facial and voice recognition to score people's trustworthiness for lenders and insurers in Japan, Indonesia, Vietnam and the Philippines. The company says the app can detect deception with 70 per cent accuracy, but researchers and privacy advocates say there is no reliable scientific basis for such judgments and warn of discrimination and privacy harms. The chief executive said the system is only one part of lenders' and insurers' decision-making and that people can choose not to use it. Critics respond that an unequal balance of power makes consent difficult.
- Company involved
- DeepScore
- AI system involved
- DeepScore
6 source articles · read the reporting →
