Plus: Instagram's CEO Adam Mosseri has denied claims that social media is "clinically addictive" AI is already making online crimes easier. It could get much worse. Just as software engineers are using artificial intelligence to help write code and check for bugs, hackers are using these tools to reduce the time and effort required to orchestrate an attack, lowering the barriers for less experienced attackers to try something out. Some in Silicon Valley warn that AI is on the brink of being able to carry out fully automated attacks. But most security researchers instead argue that we should be paying closer attention to the much more immediate risks posed by AI, which is already speeding up and increasing the volume of scams. Criminals are increasingly exploiting the latest deepfake technologies to impersonate people and swindle victims out of vast sums of money.

Scalable Vector Graphics (SVG) offer a powerful format for representing visual designs as interpretable code. Recent advances in vision-language models (VLMs) have enabled high-quality SVG generation by framing the problem as a code generation task and leveraging large-scale pretraining. VLMs are particularly suitable for this task as they capture both global semantics and fine-grained visual patterns, while transferring knowledge across vision, natural language, and code domains. However, existing VLM approaches often struggle to produce faithful and efficient SVGs because they never observe the rendered images during training. Although differentiable rendering for autoregressive SVG code generation remains unavailable, rendered outputs can still be compared to original inputs, enabling evaluative feedback suitable for reinforcement learning (RL). We introduce RLRF (Reinforcement Learning from Rendering Feedback), an RL method that enhances SVG generation in autoregressive VLMs by leveraging feedback from rendered SVG outputs. Given an input image, the model generates SVG roll-outs that are rendered and compared to the original image to compute a reward. This visual fidelity feedback guides the model toward producing more accurate, efficient, and semantically coherent SVGs. RLRF significantly outperforms supervised fine-tuning, addressing common failure modes and enabling precise, high-quality SVG generation with strong structural understanding and generalization.

Though they have improved, robots hands are still far worse than a human's. Breakthroughs, discoveries, and DIY tips sent every weekday. If you want to guess the purpose of any given futuristic humanoid robot, look at its hands. Last week, a pair of videos released by Boston Dynamics and Figure AI provided clear examples that certain tasks simply require much more "human touch." In the first case, Hyundai-owned Boston Dynamics showed off a new pair of "grippers" for its trimmed-down Atlas factory robot.

But now judges are experimenting with generative AI too. Some are confident that with the right precautions, the technology can expedite legal research, summarize cases, draft routine orders, and overall help speed up the court system, which is badly backlogged in many parts of the US. This summer, though, we've already seen AI-generated mistakes go undetected and cited by judges. A federal judge in New Jersey had to reissue an order riddled with errors that may have come from AI, and a judge in Mississippi refused to explain why his order too contained mistakes that seemed like AI hallucinations. The results of these early-adopter experiments make two things clear.

At four o'clock on a recent Friday, Kevin McCullough found himself staring at a line of text on a poster in the Graham Avenue subway station, in Williamsburg. "Prompt: What comes first, the chicken or the egg?" The poster was an ad for the School of Visual Arts. Beneath the prompt was a crude painting--of an oval-shaped chick, or was it an egg with feet and a beak?--that seemed agnostic on the issue. Something of a literalist, he had always disliked the question, believing it unworthy of endless debate.

The strength of the human hand lies in its ability to manipulate small objects precisely and robustly. In contrast, simple robotic grippers have low dexterity and fail to handle small objects effectively. This is why many automation tasks remain unsolved by robots. This paper presents an optimization-based framework for in-hand manipulation with a robotic hand equipped with compact Magnetic Tactile Sensors (MTSs). The small form factor of the robotic hand from Shadow Robot introduces challenges in estimating the state of the object while satisfying contact constraints. To address this, we formulate a trajectory optimization problem using Nonlinear Programming (NLP) for finger movements while ensuring contact points to change along the geometry of the fingers. Using the optimized trajectory from the solver, we implement and test an open-loop controller for rolling motion. To further enhance robustness and accuracy, we introduce a force controller for the fingers and a state estimator for the object utilizing MTSs. The proposed framework is validated through comparative experiments, showing that incorporating the force control with compliance consideration improves the accuracy and robustness of the rolling motion. Rolling an object with the force controller is 30\% more likely to succeed than running an open-loop controller. The demonstration video is available at https://youtu.be/6J_muL_AyE8.

Jules Rodriguez lost his voice in October of last year. His speech had been deteriorating since a diagnosis of amyotrophic lateral sclerosis (ALS) in 2020, but a tracheostomy to help him breathe dealt the final blow. Rodriguez and his wife, Maria Fernandez, who live in Miami, thought they would never hear his voice again. After feeding old recordings of Rodriguez's voice into a tool trained on voices from film, television, radio, and podcasts, the couple were able to generate a voice clone--a way for Jules to communicate in his "old voice." Rodriguez is one of over a thousand people with speech difficulties who have cloned their voices using free software from ElevenLabs.

"A tracheostomy is a scary endeavor for people living with ALS, because it signifies crossing a new stage in life, a stage that is close to the end," Rodriguez tells me using a communication device. "Before the procedure I still had some independence, and I could still speak somewhat, but now I am permanently connected to a machine that breathes for me." Rodriguez and his wife, Maria Fernandez, who live in Miami, thought they would never hear his voice again. After feeding old recordings of Rodriguez's voice into a tool trained on voices from film, television, radio, and podcasts, the couple were able to generate a voice clone--a way for Jules to communicate in his "old voice." "Hearing my voice again, after I hadn't heard it for some time, lifted my spirits," says Rodriguez, who today communicates by typing sentences using a device that tracks his eye movements, which can then be "spoken" in the cloned voice.

Tactile sensing provides robots with rich feedback during manipulation, enabling a host of perception and controls capabilities. Here, we present a new open-source, vision-based tactile sensor designed to promote reproducibility and accessibility across research and hobbyist communities. Building upon the GelSlim 3.0 sensor, our design features two key improvements: a simplified, modifiable finger structure and easily manufacturable lenses. To complement the hardware, we provide an open-source perception library that includes depth and shear field estimation algorithms to enable in-hand pose estimation, slip detection, and other manipulation tasks. Our sensor is accompanied by comprehensive manufacturing documentation, ensuring the design can be readily produced by users with varying levels of expertise. We validate the sensor's reproducibility through extensive human usability testing. For documentation, code, and data, please visit the project website: https://www.mmintlab.com/research/gelslim-4-0/

Video game performers with SAG-AFTRA will strike beginning Friday as AI "loopholes" have caused concerns. Beginning at 12:01 Friday morning, video game voice actors and motion capture performers under the Screen Actors Guild-American Federation of Television and Radio Artists will strike over artificial intelligence protections. This is the second strike for SAG-AFTRA performers in video games. While the union has conceded that wages and job safety have made gains in video game contracts, AI in interactive media continues to be a source of insecurity. TENS OF THOUSANDS OF GAMERS DESCEND ON LAS VEGAS FOR THE EVO TOURNAMENT SAG-AFTRA Chief Contracts Officer Ray Rodriguez shared at the presser on Thursday that some performers' work may be treated as "data" under current AI guidance.