As CEO of Amazon's dominant cloud computing platform AWS, Adam Selipsky is one of the most powerful people in computing at a time when the industry is racing to adopt generative artificial intelligence. Although a fan of the technology, he also has a warning for anyone trying to make sense of the moment: Some AI companies at the center of the storm are massively overhyped. Selipsky likens the generative AI rush to the early days of the dotcom bubble, when expectations spread that the internet would transform many industries almost overnight. Although in the long term the internet was indeed transformative, in the short term many projects came to nothing, and swathes of Silicon Valley companies went bust. "If you go back to, say, 1997 and you ask, 'Was the internet underhyped or overhyped?' I would argue it was underhyped," says Selipsky, who spoke with WIRED during a conference at Harvard Business School on February 4. "But if you then ask, 'Were the companies who were the leaders then dramatically overhyped?' Yes, they were."

Uyghurs outside China are traumatized. Now they're starting to talk about it The Uyghur diaspora have been forced to watch from afar as their loved ones disappear and a way of life is erased. The trauma has sparked a mental health crisis that leaders in the diaspora say is all too apparent. Many are reluctant to seek help, leaving the community's needs both underassessed and unmet. But a small group of outspoken Uyghurs is trying to change that.

The tools use generative AI to advise people whether they may be a good fit for open jobs listed on the platform and how to better tailor their profiles to stand out. The new AI features are powered by OpenAI's technology and are indicated by a sparkle emoji under job listings on LinkedIn. Clicking on it opens a chat window where a person can type queries about a job or select prewritten questions such as "Am I a good fit for this role?" Answers are provided in the form of brief bullet points sourced from scraping company profiles and other information on LinkedIn. The automated helper can also answer more specific queries about a job posting, company benefits or culture, or the industry a job is part of.

In an age where fraudsters are using generative AI to scam money or tarnish one's reputation, tech firms are coming up with methods to help users verify content -- at least still images, to begin with. As teased in its 2024 misinformation strategy, OpenAI is now including provenance metadata in images generated with ChatGPT on the web and DALL-E 3 API, with their mobile counterparts receiving the same upgrade by February 12. The metadata follows the C2PA (Coalition for Content Provenance and Authenticity) open standard, and when one such image is uploaded to the Content Credentials Verify tool, you'll be able to trace its provenance lineage. For instance, an image generated using ChatGPT will show an initial metadata manifest indicating its DALL-E 3 API origin, followed by a second metadata manifest showing that it surfaced in ChatGPT. Despite the fancy cryptographic tech behind the C2PA standard, this verification method only works when the metadata is intact; the tool is of no use if you upload an AI-generated image sans metadata -- as is the case with any screenshot or uploaded image on social media.

A New York lawyer could face discipline after it was discovered a case she cited was generated by artificial intelligence and did not actually exist. The 2nd U.S. Circuit Court of Appeals ordered lawyer Jae Lee to its grievance panel last week after discovering she used OpenAI's ChatGPT to research prior cases for a medical malpractice lawsuit but failed to confirm whether the case she was citing actually existed, according to a report from Reuters. The attorney included the fictitious state court decision in an appeal for her client's lawsuit claiming that a Queens doctor botched an abortion, according to the report, leading the court to order that Lee submit a copy of the decision that the lawyer later found she was "unable to furnish." The lawyer's conduct "falls well below the basic obligations of counsel," the 2nd U.S. Circuit Court of Appeals concluded in its disciplinary review, which was sent to Lee. Lee would later admit to using a case that was "suggested" to her by ChatGPT, a popular AI chatbot, and failing to verify the results herself. The lawyer's decision to use the popular application comes even though experts have warned against such practices, noting that AI is a relatively new technology that also is well-known for "hallucinating" false or misleading results.

Counterfactual image editing is an important task in generative AI, which asks how an image would look if certain features were different. The current literature on the topic focuses primarily on changing individual features while remaining silent about the causal relationships between these features, as present in the real world. In this paper, we formalize the counterfactual image editing task using formal language, modeling the causal relationships between latent generative factors and images through a special type of model called augmented structural causal models (ASCMs). Second, we show two fundamental impossibility results: (1) counterfactual editing is impossible from i.i.d. image samples and their corresponding labels alone; (2) even when the causal relationships between the latent generative factors and images are available, no guarantees regarding the output of the model can be provided. Third, we propose a relaxation for this challenging problem by approximating non-identifiable counterfactual distributions with a new family of counterfactual-consistent estimators. This family exhibits the desirable property of preserving features that the user cares about across both factual and counterfactual worlds. Finally, we develop an efficient algorithm to generate counterfactual images by leveraging neural causal models.

Deep Generative Models (DGMs) have been shown to be powerful tools for generating tabular data, as they have been increasingly able to capture the complex distributions that characterize them. However, to generate realistic synthetic data, it is often not enough to have a good approximation of their distribution, as it also requires compliance with constraints that encode essential background knowledge on the problem at hand. In this paper, we address this limitation and show how DGMs for tabular data can be transformed into Constrained Deep Generative Models (C-DGMs), whose generated samples are guaranteed to be compliant with the given constraints. This is achieved by automatically parsing the constraints and transforming them into a Constraint Layer (CL) seamlessly integrated with the DGM. Our extensive experimental analysis with various DGMs and tasks reveals that standard DGMs often violate constraints, some exceeding $95\%$ non-compliance, while their corresponding C-DGMs are never non-compliant. Then, we quantitatively demonstrate that, at training time, C-DGMs are able to exploit the background knowledge expressed by the constraints to outperform their standard counterparts with up to $6.5\%$ improvement in utility and detection. Further, we show how our CL does not necessarily need to be integrated at training time, as it can be also used as a guardrail at inference time, still producing some improvements in the overall performance of the models. Finally, we show that our CL does not hinder the sample generation time of the models.

This study consists of a novel approach toward the analysis of court judgments spanning five countries, including the United States, the United Kingdom, Rwanda, Sweden and Hong Kong. This study also explores the intersection of the latest advancements in artificial intelligence (AI) and legal analysis, emphasizing the role of AI (specifically generative AI) in identifying human biases and facilitating automated, valid, and coherent multisided argumentation of court judgments with the goal of ensuring consistent application of laws in and across various jurisdictions. By incorporating Advanced Language Models (ALMs) and a newly introduced human-AI collaborative framework, this paper seeks to analyze Grounded Theory-based research design with Advanced Language Models (ALMs) in the practice of law. SHIRLEY is the name of the AI-based application (built on top of OpenAI's GPT technology), focusing on detecting logical inconsistencies and biases across various legal decisions. SHIRLEY analysis is aggregated and is accompanied by a comparison-oriented AI-based application called SAM (also an ALM) to identify relative deviations in SHIRLEY bias detections. Further, a CRITIC is generated within semi-autonomous arbitration process via the ALM, SARA. A novel approach is introduced in the utilization of an AI arbitrator to critically evaluate biases and qualitative-in-nature nuances identified by the aforementioned AI applications (SAM in concert with SHIRLEY), based on the Hague Rules on Business and Human Rights Arbitration. This Semi-Automated Arbitration Process (SAAP) aims to uphold the integrity and fairness of legal judgments by ensuring a nuanced debate-resultant "understanding" through a hybrid system of AI and human-based collaborative analysis.

The rapid advancement of Large AI Models (LAIMs), particularly diffusion models and large language models, has marked a new era where AI-generated multimedia is increasingly integrated into various aspects of daily life. Although beneficial in numerous fields, this content presents significant risks, including potential misuse, societal disruptions, and ethical concerns. Consequently, detecting multimedia generated by LAIMs has become crucial, with a marked rise in related research. Despite this, there remains a notable gap in systematic surveys that focus specifically on detecting LAIM-generated multimedia. Addressing this, we provide the first survey to comprehensively cover existing research on detecting multimedia (such as text, images, videos, audio, and multimodal content) created by LAIMs. Specifically, we introduce a novel taxonomy for detection methods, categorized by media modality, and aligned with two perspectives: pure detection (aiming to enhance detection performance) and beyond detection (adding attributes like generalizability, robustness, and interpretability to detectors). Additionally, we have presented a brief overview of generation mechanisms, public datasets, and online detection tools to provide a valuable resource for researchers and practitioners in this field. Furthermore, we identify current challenges in detection and propose directions for future research that address unexplored, ongoing, and emerging issues in detecting multimedia generated by LAIMs. Our aim for this survey is to fill an academic gap and contribute to global AI security efforts, helping to ensure the integrity of information in the digital realm. The project link is https://github.com/Purdue-M2/Detect-LAIM-generated-Multimedia-Survey.

Generative adversarial networks (GANs) usually struggle in learning from highly diverse data, whose underlying manifold is complex. In this work, we revisit the mathematical foundations of GANs, and theoretically reveal that the native adversarial loss for GAN training is insufficient to fix the problem of subsets with positive Lebesgue measure of the generated data manifold lying out of the real data manifold. Instead, we find that score matching serves as a promising solution to this issue thanks to its capability of persistently pushing the generated data points towards the real data manifold. We thereby propose to improve the optimization of GANs with score matching regularity (SMaRt). Regarding the empirical evidences, we first design a toy example to show that training GANs by the aid of a ground-truth score function can help reproduce the real data distribution more accurately, and then confirm that our approach can consistently boost the synthesis performance of various state-of-the-art GANs on real-world datasets with pre-trained diffusion models acting as the approximate score function. For instance, when training Aurora on the ImageNet 64x64 dataset, we manage to improve FID from 8.87 to 7.11, on par with the performance of one-step consistency model. The source code will be made public.