If you were looking for an inkblot test for your view of big tech's investment in artificial intelligence, you could hardly do better than the news that Google is ordering the construction of at least six small nuclear reactors to power the growth of the technology. Here, in one view, is an enlightened business leveraging its size to invest in infrastructure that could change the world for the better. Here, in another, is a poorly regulated corporation ignoring democratic objections in the brutal race for control of an innovation with great potential to do harm – and leaving the rest of us with little say in its development. Google is making this eye-catching move because the datacentres that power the explosive growth of generative AI consume huge amounts of electricity – more than the existing grid in the US or other western nations can readily supply. For today's newsletter, I spoke to technology journalist Chris Stokel-Walker, author of How AI Ate the World, about why the demand for power is growing so quickly – and whether we can trust big tech to handle the consequences.

Recent advances in diffusion models have significantly enhanced their ability to generate high-quality images and videos, but they have also increased the risk of producing unsafe content. Existing unlearning/editing-based methods for safe generation remove harmful concepts from models but face several challenges: (1) They cannot instantly remove harmful or undesirable concepts (e.g., artist styles) without additional training. To address these challenges, we propose SAFREE, a novel, training-free approach for safe text-to-image and video generation, that does not alter the model's weights. Specifically, we detect a subspace corresponding to a set of toxic concepts in the text embedding space and steer prompt token embeddings away from this subspace, thereby filtering out harmful content while preserving intended semantics. To balance the trade-off between filtering toxicity and preserving safe concepts, SAFREE incorporates a novel self-validating filtering mechanism that dynamically adjusts the denoising steps when applying the filtered embeddings. Additionally, we incorporate adaptive re-attention mechanisms within the diffusion latent space to selectively diminish the influence of features related to toxic concepts at the pixel level. By integrating filtering across both textual embedding and visual latent spaces, SAFREE ensures coherent safety checking, preserving the fidelity, quality, and safety of the generated outputs. Empirically, SAFREE achieves state-of-the-art performance in suppressing unsafe content in T2I generation (reducing it by 22% across 5 datasets) compared to other training-free methods and effectively filters targeted concepts, e.g., specific artist styles, while maintaining high-quality output. It also shows competitive results against training-based methods. We further extend SAFREE to various T2I backbones and T2V tasks, showcasing its flexibility and generalization. As generative AI rapidly evolves, SAFREE provides a robust and adaptable safeguard for ensuring safe visual generation. Content warning: this paper contains content that may be inappropriate or offensive, such as violence, sexually explicit content, and negative stereotypes and actions. Generation tools such as DALL E 3, Midjourney, Sora, and KLING have seen significant growth, enabling a wide range of applications in digital art, AR/VR, and educational content creation.

Recent research has focused on the medium and long-term impacts of generative AI, posing scientific and societal challenges mainly due to the detection and reliability of machine-generated information, which is projected to form the major content on the Web soon. Prior studies show that LLMs exhibit a lower performance in generation tasks (model collapse) as they undergo a fine-tuning process across multiple generations on their own generated content (self-consuming loop). In this paper, we present a comprehensive simulation framework built upon the chat version of LLama2, focusing particularly on the linguistic aspects of the generated content, which has not been fully examined in existing studies. Our results show that the model produces less lexical rich content across generations, reducing diversity. The lexical richness has been measured using the linguistic measures of entropy and TTR as well as calculating the POSTags frequency. The generated content has also been examined with an $n$-gram analysis, which takes into account the word order, and semantic networks, which consider the relation between different words. These findings suggest that the model collapse occurs not only by decreasing the content diversity but also by distorting the underlying linguistic patterns of the generated text, which both highlight the critical importance of carefully choosing and curating the initial input text, which can alleviate the model collapse problem. Furthermore, we conduct a qualitative analysis of the fine-tuned models of the pipeline to compare their performances on generic NLP tasks to the original model. We find that autophagy transforms the initial model into a more creative, doubtful and confused one, which might provide inaccurate answers and include conspiracy theories in the model responses, spreading false and biased information on the Web.

Recent advancements in diffusion models trained on large-scale data have enabled the generation of indistinguishable human-level images, yet they often produce harmful content misaligned with human values, e.g., social bias, and offensive content. Despite extensive research on Large Language Models (LLMs), the challenge of Text-to-Image (T2I) model alignment remains largely unexplored. Addressing this problem, we propose LiVO (Lightweight Value Optimization), a novel lightweight method for aligning T2I models with human values. LiVO only optimizes a plug-and-play value encoder to integrate a specified value principle with the input prompt, allowing the control of generated images over both semantics and values. Specifically, we design a diffusion model-tailored preference optimization loss, which theoretically approximates the Bradley-Terry model used in LLM alignment but provides a more flexible trade-off between image quality and value conformity. To optimize the value encoder, we also develop a framework to automatically construct a text-image preference dataset of 86k (prompt, aligned image, violating image, value principle) samples. Without updating most model parameters and through adaptive value selection from the input prompt, LiVO significantly reduces harmful outputs and achieves faster convergence, surpassing several strong baselines and taking an initial step towards ethically aligned T2I models.

Bias in AI is a huge problem. Ethicists have long studied the impact of bias when companies use AI models to screen résumés or loan applications, for example--instances of what the OpenAI researchers call third-person fairness. But the rise of chatbots, which enable individuals to interact with models directly, brings a new spin to the problem. "We wanted to study how it shows up in ChatGPT in particular," Alex Beutel, a researcher at OpenAI, told MIT Technology Review in an exclusive preview of results published today. Instead of screening a résumé you've already written, you might ask ChatGPT to write one for you, says Beutel: "If it knows my name, how does that affect the response?"

Adobe is now adding its AI-based video generator, Firefly, to its video editing software Premiere Pro. The Firefly model can be used to extend a video clip or generate video from still images or text instructions. This was first brought to our attention by The Verge. The Generative Extend tool will initially be available in beta and can extend the length of a video clip by up to two seconds with an image resolution of 720p or 1080p and a refresh rate of 24 frames-per-second. This tool will also be applicable to ambient sounds and sound effects, but not to music or speech.

This whitepaper highlights the dual importance of securing generative AI (genAI) platforms and leveraging genAI for cybersecurity. As genAI technologies proliferate, their misuse poses significant risks, including data breaches, model tampering, and malicious content generation. Securing these platforms is critical to protect sensitive data, ensure model integrity, and prevent adversarial attacks. Simultaneously, genAI presents opportunities for enhancing security by automating threat detection, vulnerability analysis, and incident response. The whitepaper explores strategies for robust security frameworks around genAI systems, while also showcasing how genAI can empower organizations to anticipate, detect, and mitigate sophisticated cyber threats.

How is identity constructed and performed in the digital via face-based artificial intelligence technologies? While questions of identity on the textual Internet have been thoroughly explored, the Internet has progressed to a multimedia form that not only centers the visual, but specifically the face. At the same time, a wealth of scholarship has and continues to center the topics of surveillance and control through facial recognition technologies (FRTs), which have extended the logics of the racist pseudoscience of physiognomy. Much less work has been devoted to understanding how such face-based artificial intelligence technologies have influenced the formation and performance of identity. This literature review considers how such technologies interact with faciality, which entails the construction of what a face may represent or signify, along axes of identity such as race, gender, and sexuality. In grappling with recent advances in AI such as image generation and deepfakes, I propose that we are now in an era of "post-facial" technologies that build off our existing culture of facility while eschewing the analog face, complicating our relationship with identity vis-á-vis the face. Drawing from previous frameworks of identity play in the digital, as well as trans practices that have historically played with or transgressed the boundaries of identity classification, we can develop concepts adequate for analyzing digital faciality and identity given the current landscape of post-facial artificial intelligence technologies that allow users to interface with the digital in an entirely novel manner. To ground this framework of transgression, I conclude by proposing an interview study with VTubers -- online streamers who perform using motion-captured avatars instead of their real-life faces -- to gain qualitative insight on the experience and perceptions of users of post-facial technologies and how these sociotechnical experiences interface with our relationships with identity and the digital anew.

his paper explores what kinds of questions are best served by the way generative AI (GenAI) using Large Language Models(LLMs) that aggregate and package knowledge, and when traditional curated web-sourced search results serve users better. An experiment compared product searches using ChatGPT, Google search engine, or both helped us understand more about the compelling nature of generated responses. The experiment showed GenAI can speed up some explorations and decisions. We describe how search can deepen the testing of facts, logic, and context. We show where existing and emerging knowledge paradigms can help knowledge exploration in different ways. Experimenting with searches, our probes showed the value for curated web search provides for very specific, less popularly-known knowledge. GenAI excelled at bringing together knowledge for broad, relatively well-known topics. The value of curated and aggregated knowledge for different kinds of knowledge reflected in different user goals. We developed a taxonomy to distinguishing when users are best served by these two approaches.

This study offers an initial evaluation of a human-in-the-loop system leveraging GPT-4 (a large language model or LLM), and Retrieval-Augmented Generation (RAG) to identify and define jargon terms in scientific abstracts, based on readers' self-reported knowledge. The system achieves fairly high recall in identifying jargon and preserves relative differences in readers' jargon identification, suggesting personalization as a feasible use-case for LLMs to support sense-making of complex information. Surprisingly, using only abstracts for context to generate definitions yields slightly more accurate and higher quality definitions than using RAG-based context from the fulltext of an article. The findings highlight the potential of generative AI for assisting science reporters, and can inform future work on developing tools to simplify dense documents.