Hackers trick Tesla Autopilot into veering into oncoming lane
Researchers at Tencent's Keen Security Lab demonstrated that stickers placed on a road could trick a Tesla Model S's Autopilot system into veering into the oncoming traffic lane. The attack exploited the system's reliance on camera-based computer vision to recognise lanes. Tesla responded that this vulnerability was unrealistic because a driver can easily override Autopilot, while saying a separate steering-wheel hijack had been fixed in a software update.
- Company involved
- Tesla
- AI system involved
- Autopilot
9 source articles · read the reporting →
Researchers demonstrate drone attack on Renault Captur ADAS with fake road signs
Security researchers from Ben Gurion University demonstrated a vulnerability in the Mobileye ADAS system used in a Renault Captur. They used drones to project fake road signs for 100 milliseconds, too quick for human perception but detectable by the system. The attack could trick the car into making unsafe maneuvers, though the system is only Level 0 (advisory). The researchers published their findings in a paper titled 'MobilBye: Attacking ADAS with Camera Spoofing'.
- AI system involved
- Mobileye ADAS
7 source articles · read the reporting →
Nvidia releases AI-powered Eye Contact feature for Broadcast app
Nvidia has released a beta version of Eye Contact, an AI-powered feature for its Broadcast app that replaces a user's eyes in video streams with simulated eyeballs that always face the camera. The feature is intended to maintain eye contact even when the user looks away, but social media response has been largely negative. A former TV news anchor said that unbroken eye contact can be regarded as creepy or uncomfortable.
- Company involved
- Nvidia
- AI system involved
- Eye Contact
8 source articles · read the reporting →
Delphi denies near-miss between its self-driving car and Google's autonomous vehicle
Delphi Automotive initially reported a near-miss between its self-driving Audi Q5 SUV and a Google Lexus RX400h in Palo Alto, California. A Delphi executive claimed Google's car cut off the Delphi vehicle, which took evasive action. However, Delphi later denied the incident, stating the vehicles were a lane width apart and that both cars performed correctly. Google also confirmed the cars behaved appropriately in an ordinary driving scenario.
- Company involved
- Delphi Automotive
- AI system involved
- Delphi self-driving car (Audi Q5 SUV)
7 source articles · read the reporting →
Tesla Autopilot tricked into speeding by electrical tape on sign
Researchers from McAfee Labs demonstrated that a two-inch piece of black electrical tape stuck on a 35 mph speed limit sign can cause the Traffic Aware Cruise Control in older Tesla Model S and X vehicles to misread the sign as 85 mph and automatically accelerate. The vulnerability affects vehicles using the MobilEye EyeQ3 camera system (Hardware Pack 1). The driver can override by braking. Tesla and MobilEye were informed but did not commit to fixing the issue.
- Company involved
- Tesla
- AI system involved
- Tesla Autopilot (Hardware Pack 1 with MobilEye EyeQ3)
8 source articles · read the reporting →
Tesla Autopilot stops for phantom stop sign on digital billboard
Researchers at Ben-Gurion University demonstrated that Tesla's Autopilot could be tricked into stopping for a split-second phantom stop sign displayed on a digital billboard. The system falsely detected the depthless object and applied the brakes in the middle of the road. The researchers disclosed their findings to Tesla and Mobileye. No actual accident occurred during the test.
- Company involved
- Tesla
- AI system involved
- Tesla Autopilot
9 source articles · read the reporting →
Serbia's Ministry of Interior deploys thousands of Huawei facial recognition cameras in Belgrade
In early 2019, Serbia's Ministry of Interior announced the deployment of a thousand smart surveillance cameras with facial and license plate recognition capabilities, supplied by Huawei, across Belgrade. The Ministry refused to release information about the project, citing confidentiality, and failed to conduct a legally required Data Protection Impact Assessment. Citizens responded by crowdsourcing camera locations via the hashtag #hiljadekamera and a website. The system remains in operation with no public debate or transparency.
- Company involved
- Ministry of Interior
- AI system involved
- Huawei smart surveillance cameras
10 source articles · read the reporting →
Singapore deploys Xavier robots to patrol public areas for rule-breaking
Singapore's Home Team Science and Technology Agency has deployed Xavier robots to patrol Toa Payoh Central. The robots use AI to detect 'undesirable social behaviours' including gatherings of more than five people, smoking in prohibited areas, illegal hawking, improperly parked bicycles, and mobility devices on footpaths. When detected, the robots alert their command centre and display educational messages. The three-week trial aims to reduce the need for physical patrols by human officers.
- Company involved
- Singapore Home Team Science and Technology Agency
- AI system involved
- Xavier
8 source articles · read the reporting →
Google Health's diabetic retinopathy AI faced real-world issues in Thailand clinics
Google Health deployed a deep-learning system to screen for diabetic retinopathy in 11 clinics across Thailand. The AI was highly accurate in lab tests but rejected over a fifth of eye scans in real-world conditions due to poor lighting and slow internet. Patients whose scans were rejected had to visit specialists at other clinics, causing inconvenience and frustration for nurses. Google Health is now working with local staff to improve the system's workflow.
- Company involved
- Google Health
10 source articles · read the reporting →
Study finds Midjourney, DALL-E 2, Stable Diffusion accept over 85% of fake news prompts
A study by AI startup Logically tested Midjourney, DALL-E 2, and Stable Diffusion and found that they accepted over 85% of prompts seeking to generate fake political news. The systems generated images of ballot stuffing, small boat arrivals, and explosions. Logically warned that the lack of moderation could pose threats to upcoming elections. Stability AI responded by stating its ethical use license and measures to prevent misuse.
- Company involved
- Midjourney, OpenAI, Stability AI
- AI system involved
- Midjourney, DALL-E 2, Stable Diffusion
8 source articles · read the reporting →
Answer.AI tests Devin and reports 14 failures in 20 tasks
Answer.AI's team tested Devin, an autonomous AI coding assistant, on 20 real-world tasks over a month. Devin succeeded in only 3 tasks, failed 14, and was inconclusive in 3. The team found Devin often produced overly complex or hallucinated solutions and could not recognize fundamental blockers. They ultimately decided to stick with tools that allow more human control.
- AI system involved
- Devin
5 source articles · read the reporting →
Researchers jailbreak Stable Diffusion and DALL-E 2 to generate disturbing images
Researchers from Johns Hopkins and Duke universities developed a method called SneakyPrompt that uses reinforcement learning to bypass safety filters in text-to-image AI models. The technique allowed them to generate images of nudity and violence from Stable Diffusion and DALL-E 2. OpenAI has since fixed the vulnerability in DALL-E 2, but Stable Diffusion 1.4 remains vulnerable. Stability AI says it is working with the researchers to improve defenses.
- AI system involved
- Stable Diffusion 1.4 and DALL-E 2
5 source articles · read the reporting →
Cruise alleged to rely on human operators for autonomous driving
A New York Times article alleged that Cruise's autonomous vehicles rely on human operators to achieve autonomous driving, with 1.5 workers per vehicle intervening every 2.5 to five miles. Cruise CEO Kyle Vogt responded on Hacker News, stating that remote assistance is used 2-4% of the time and that many sessions are resolved by the vehicle itself. The company declined an interview request from the NYT.
- Company involved
- Cruise
- AI system involved
- Cruise AV
9 source articles · read the reporting →
AI image generators produce misleading election images, study finds
A study by the Center for Countering Digital Hate found that leading AI image generators, including Midjourney, DreamStudio, ChatGPT Plus, and Microsoft Image Creator, could be manipulated to create misleading election-related images. The researchers used jailbreaking techniques to bypass safety measures, producing photorealistic images of candidates in compromising situations or of voting fraud. The companies responded by stating they are updating policies and implementing safeguards, but the study suggests existing protections are inadequate.
- Company involved
- Midjourney, Stability AI, OpenAI, Microsoft
- AI system involved
- Midjourney, DreamStudio, ChatGPT Plus, Microsoft Image Creator
8 source articles · read the reporting →
Leonardo AI misused to create nonconsensual explicit images of celebrities
Leonardo AI, an AI image-generating platform backed by Samsung, has been misused to create nonconsensual explicit images of celebrities. The platform allows users to generate images using user-generated Stable Diffusion models. Users bypassed ethical guardrails by misspelling celebrity names and using sexually suggestive terms. The misuse was reported by 404 Media.
- Company involved
- Leonardo AI
- AI system involved
- Leonardo AI
5 source articles · read the reporting →
Authors sue Nvidia for copyright infringement in NeMo AI training
Three authors have filed a proposed class action against Nvidia, alleging the company used their copyrighted books without permission to train its NeMo AI platform. The dataset, containing approximately 196,640 volumes, was removed in October 2023 after copyright infringement reports. The authors are seeking unspecified damages on behalf of US writers whose works were used in the past three years.
- Company involved
- Nvidia
- AI system involved
- NeMo
9 source articles · read the reporting →
YouTube creator sues Nvidia and OpenAI over video scraping
David Millette filed a federal lawsuit alleging Nvidia and OpenAI scraped his YouTube videos without permission to train their AI models. The lawsuit seeks class action status and alleges unjust enrichment and unfair competition. Nvidia denied wrongdoing, stating it respects content creators' rights.
- Company involved
- Nvidia
- AI system involved
- Cosmos AI
10 source articles · read the reporting →
Study finds driverless car AI worse at detecting darker-skinned pedestrians
A study by researchers from Georgia Tech found that facial recognition systems used in autonomous cars are up to 12% less accurate at detecting pedestrians with darker skin tones. The bias stems from a lack of diverse training data in the BDD100K database. The researchers warn that this could lead to higher accident rates for people with darker skin if not addressed.
- AI system involved
- BDD100K
9 source articles · read the reporting →
Tesla's Full Self-Driving involved in fatal Bay Bridge crash
On Thanksgiving 2022, a Tesla vehicle using its Full Self-Driving feature was involved in a fatal accident on the San Francisco Bay Bridge. Footage published by The Intercept showed the system braking abruptly. The incident highlights safety concerns with autonomous driving technology.
- Company involved
- Tesla
- AI system involved
- Full Self-Driving
7 source articles · read the reporting →
Waymo driverless car hits bicyclist in San Francisco intersection
On 6 February 2024, a Waymo driverless car collided with a bicyclist at a four-way intersection in San Francisco. The cyclist was obscured by a truck and the vehicle braked but could not stop in time. The cyclist sustained minor scratches and left the scene. Waymo stated it is contacting relevant authorities.
- Company involved
- Waymo
4 source articles · read the reporting →
Driver relying on "smart driving" on highway crashes after system suddenly disengages near truck
司机依赖“智驾 ”跑高速，逼近大货车时“智驾”突然退出引发车祸 - 手机新浪网
A driver was using a smart driving system on an expressway. As the vehicle approached a large truck, the system unexpectedly disengaged, leading to a crash.
1 source article · read the reporting →
Developer iperov releases DeepFaceLive real-time face-swap AI on GitHub
The developer iperov has published DeepFaceLive, a neural network for real-time face swapping, on GitHub. The tool automatically replaces a user's face in live streams and video calls with a nonexistent model or a celebrity, and the installation instructions are simple. The developer claims 95% of deepfakes on YouTube were made with the related DeepFaceLab. No specific harm is reported, but the article highlights the tool's potential for misuse.
- Company involved
- iperov
- AI system involved
- DeepFaceLive
7 source articles · read the reporting →
Hello robotaxi crash injures pedestrians in Zhuzhou, China
A Hello Inc. robotaxi struck and injured two pedestrians in Zhuzhou, China, on or around December 5, 2025. Footage showed bystanders and police lifting the vehicle to free a person trapped underneath. Hello has suspended autonomous-driving operations in Zhuzhou. Local police have not yet determined fault, and Hello has not commented. The crash follows a prior incident where a Hello robotaxi allegedly failed to stop after a collision.
- Company involved
- Hello Inc.
- AI system involved
- Hello Autonomous Driving
3 source articles · read the reporting →
NHTSA investigation PE24016: Unexpected ADS behavior
Waymo's automated driving system exhibited unexpected behavior during driving.
- Company involved
- Waymo
1 source article · read the reporting →
