Image-generating AI models like DALL-E 2 and Stable Diffusion can -- and do -- replicate aspects of images from their training data, researchers show in a new study, raising concerns as these services enter wide commercial use. The study hasn't been peer reviewed yet, and the co-authors submitted it to a conference whose rules forbid media interviews until the research has been accepted for publication. But one of the researchers, who asked not to be identified by name, shared high-level thoughts with TechCrunch via email. "Even though diffusion models such as Stable Diffusion produce beautiful images, and often ones that appear highly original and custom tailored to a particular text prompt, we show that these images may actually be copied from their training data, either wholesale or by copying only parts of training images," the researcher said. "Companies generating data with diffusion models may need to reconsider wherever intellectual property laws are concerned. It is virtually impossible to verify that any particular image generated by Stable Diffusion is novel and not stolen from the training set."

Motivated by mitigating potentially harmful impacts of technologies, the AI community has formulated and accepted mathematical definitions for certain pillars of accountability: e.g. privacy, fairness, and model transparency. Yet, we argue this is fundamentally misguided because these definitions are imperfect, siloed constructions of the human values they hope to proxy, while giving the guise that those values are sufficiently embedded in our technologies. Under popularized methods, tensions arise when practitioners attempt to achieve each pillar of fairness, privacy, and transparency in isolation or simultaneously. In this position paper, we push for redirection. We argue that the AI community needs to consider all the consequences of choosing certain formulations of these pillars -- not just the technical incompatibilities, but also the effects within the context of deployment. We point towards sociotechnical research for frameworks for the latter, but push for broader efforts into implementing these in practice.

Scaling up neural networks has led to remarkable performance across a wide range of tasks. Moreover, performance often follows reliable scaling laws as a function of training set size, model size, and compute, which offers valuable guidance as large-scale experiments are becoming increasingly expensive. However, previous work on scaling laws has primarily used private data \& models or focused on uni-modal language or vision learning. To address these limitations, we investigate scaling laws for contrastive language-image pre-training (CLIP) with the public LAION dataset and the open-source OpenCLIP repository. Our large-scale experiments involve models trained on up to two billion image-text pairs and identify power law scaling for multiple downstream tasks including zero-shot classification, retrieval, linear probing, and end-to-end fine-tuning. We find that the training distribution plays a key role in scaling laws as the OpenAI and OpenCLIP models exhibit different scaling behavior despite identical model architectures and similar training recipes. We open-source our evaluation workflow and all models, including the largest public CLIP models, to ensure reproducibility and make scaling laws research more accessible. Source code and instructions to reproduce this study will be available at https://github.com/LAION-AI/scaling-laws-openclip

Complex systems often contain feedback loops that can be described as cyclic causal models. Intervening in such systems may lead to counterintuitive effects, which cannot be inferred directly from the graph structure. After establishing a framework for differentiable soft interventions based on Lie groups, we take advantage of modern automatic differentiation techniques and their application to implicit functions in order to optimize interventions in cyclic causal models. We illustrate the use of this framework by investigating scenarios of transition to sustainable economies.

Advances in image processing and analysis as well as machine learning techniques have contributed to the use of biometric recognition systems in daily people tasks. These tasks range from simple access to mobile devices to tagging friends in photos shared on social networks and complex financial operations on self-service devices for banking transactions. In China, the use of these systems goes beyond personal use becoming a country's government policy with the objective of monitoring the behavior of its population. On July 05th 2021, the Brazilian government announced acquisition of a biometric recognition system to be used nationwide. In the opposite direction to China, Europe and some American cities have already started the discussion about the legality of using biometric systems in public places, even banning this practice in their territory. In order to open a deeper discussion about the risks and legality of using these systems, this work exposes the vulnerabilities of biometric recognition systems, focusing its efforts on the face modality. Furthermore, it shows how it is possible to fool a biometric system through a well-known presentation attack approach in the literature called morphing. Finally, a list of ten concerns was created to start the discussion about the security of citizen data and data privacy law in the Age of Artificial Intelligence (AI).

The costs and impacts of government corruption range from impairing a country's economic growth to affecting its citizens' well-being and safety. Public contracting between government dependencies and private sector instances, referred to as public procurement, is a fertile land of opportunity for corrupt practices, generating substantial monetary losses worldwide. Thus, identifying and deterring corrupt activities between the government and the private sector is paramount. However, due to several factors, corruption in public procurement is challenging to identify and track, leading to corrupt practices going unnoticed. This paper proposes a machine learning model based on an ensemble of random forest classifiers, which we call hyper-forest, to identify and predict corrupt contracts in M\'exico's public procurement data. This method's results correctly detect most of the corrupt and non-corrupt contracts evaluated in the dataset. Furthermore, we found that the most critical predictors considered in the model are those related to the relationship between buyers and suppliers rather than those related to features of individual contracts. Also, the method proposed here is general enough to be trained with data from other countries. Overall, our work presents a tool that can help in the decision-making process to identify, predict and analyze corruption in public procurement contracts.

China's Cyberspace Administration recently issued regulations prohibiting the creation of AI-generated media without clear labels, such as watermarks--among other policies--reports The Register. The new rules come as part of China's evolving response to the generative AI trend that has swept the tech world in 2022, and they will take effect on January 10, 2023. In China, the Cyberspace Administration oversees the regulation, oversight, and censorship of the Internet. Under the new regulations, the administration will keep a closer eye on what it calls "deep synthesis" technology. In a news post on the website of China's Office of the Central Cyberspace Affairs Commission, the government outlined its reasons for issuing the regulation.

The bold, hype-laden pronouncements around AI and machine learning were hard to miss five or six years ago. Headlines about robo-accountants stealing jobs, algorithms that will cure disease and autonomous vehicles were everywhere. Then, reality quickly caught up with the hype, those promises eventually proved overly ambitious and many people lost the plot. Seemingly all at once, just as most of the mainstream media was starting to write off AI, the pendulum swung in the other direction. The last two years have seen a surge in AI adoption.

When Melissa Heikkilä, our senior AI reporter, tried the new viral AI avatar app Lensa, she was hoping to get results similar to other colleagues at MIT Technology Review, who got realistic yet flattering avatars--think astronauts, and fierce warriors. Instead, she got tons of nudes. Out of the generated 100 avatars, 16 were topless, while another 14 depicted her in extremely skimpy clothes and overtly sexualized poses. Many of the avatars were of generic Asian women clearly modeled on anime or video-game characters, or, most likely, porn. Another colleague with Chinese heritage got similar results: reams and reams of pornified avatars. Its results are generated using Stable Diffusion, an AI model that draws from a massive open-source data set compiled by scraping images from the internet.

The question of whether artificial intelligence can invent is nearly 200 years old, going back to the very beginning of computing. Victorian mathematician Ada Lovelace wrote what's generally considered the first computer programme. As she did, she wondered about the limits of what computers could do. The Analytical Engine has no pretensions to originate anything. It can do whatever we know how to order it to perform.