The Supreme Court's conservative justices want to reduce government regulation of private industry. Of that they are certain. Unless the private industry is artificial intelligence or social media. On that, they need some time to think. This summer, the Supreme Court issued two decisions that will impact the future of A.I. regulation.

A California bill that would establish first-in-the-nation safety measures for the largest artificial intelligence systems cleared an important vote Wednesday. The proposal, aiming to reduce potential risks created by AI, would require companies to test their models and publicly disclose their safety protocols to prevent the models from being manipulated to, for example, wipe out the state's electric grid or help build chemical weapons – scenarios experts say could be possible in the future with such rapid advancements in the industry. The measure squeaked by in the state assembly Wednesday and now faces a final vote in the state senate, where it has passed once already, before it heads to the governor's desk for his signature, though he has not indicated his position on it. Governor Gavin Newsom then has until the end of September to decide whether to sign it into law, veto it or allow it to become law without his signature. He declined to weigh in on the measure earlier this summer but had warned against AI overregulation.

Cate Blanchett – beloved thespian, film star and refugee advocate – is standing at a lectern, addressing the European Union parliament. "The future is now," she says, authoritatively. So far, so normal, until: "But where the fuck are the sex robots?" The footage is from a 2023 address that Blanchett actually gave – but the rest has been made up. Her voice was generated by Australian artist Xanthe Dobbie using the text-to-speech platform PlayHT, for Dobbie's 2024 video work Future Sex/Love Sounds – an imagining of a sex robot-induced feminist utopia, voiced by celebrity clones.

When news broke that AI-generated nude pictures of students were popping up at a Beverly Hills Middle School in February, many district officials and parents were horrified. But others said no one should have been blindsided by the spread of AI-powered "undressing" programs. "The only thing shocking about this story," one Carlsbad parent said his 14-year-old told him, "is that people are shocked." Now, a newly released report by Thorn, a tech company that works to stop the spread of child sexual abuse material, shows how common deepfake abuse has become. The proliferation coincides with the wide availability of cheap "undressing" apps and other easy-to-use, AI-powered programs to create deepfake nudes.

Google has officially been named a monopoly. On Aug. 5, a federal judge charged the tech giant with illegally using its market power to harm rival search engines, marking the first antitrust defeat for a major internet platform in more than 20 years--and thereby calling into question the business practices of Silicon Valley's most powerful companies. Many experts have speculated the landmark decision will make judges more receptive to antitrust action in other ongoing cases against the Big Tech platforms, especially with regards to the burgeoning AI industry. Today, the AI ecosystem is dominated by many of the same companies that the government is challenging in court, and those companies are using the same tactics to entrench their power in AI markets. Judge Amit Mehta's ruling in the Google case centered on the massive sums of money the company paid firms like Apple and Samsung to make its search engine the default on their smartphones and browsers.

While Large Language Models (LLMs) excel in text generation and question-answering, their effectiveness in AI legal and policy is limited by outdated knowledge, hallucinations, and inadequate reasoning in complex contexts. Retrieval-Augmented Generation (RAG) systems improve response accuracy by integrating external knowledge but struggle with retrieval errors, poor context integration, and high costs, particularly in interpreting qualitative and quantitative AI legal texts. This paper introduces a Hybrid Parameter-Adaptive RAG (HyPA-RAG) system tailored for AI legal and policy, exemplified by NYC Local Law 144 (LL144). HyPA-RAG uses a query complexity classifier for adaptive parameter tuning, a hybrid retrieval strategy combining dense, sparse, and knowledge graph methods, and an evaluation framework with specific question types and metrics. By dynamically adjusting parameters, HyPA-RAG significantly improves retrieval accuracy and response fidelity. Testing on LL144 shows enhanced correctness, faithfulness, and contextual precision, addressing the need for adaptable NLP systems in complex, high-stakes AI legal and policy applications.

As foundation models have accumulated hundreds of millions of users, developers have begun to take steps to prevent harmful types of uses. One salient intervention that foundation model developers adopt is acceptable use policies: legally binding policies that prohibit users from using a model for specific purposes. This paper identifies acceptable use policies from 30 foundation model developers, analyzes the use restrictions they contain, and argues that acceptable use policies are an important lens for understanding the regulation of foundation models. Taken together, developers' acceptable use policies include 127 distinct use restrictions; the wide variety in the number and type of use restrictions may create fragmentation across the AI supply chain. Developers also employ acceptable use policies to prevent competitors or specific industries from making use of their models. Developers alone decide what constitutes acceptable use, and rarely provide transparency about how they enforce their policies. In practice, acceptable use policies are difficult to enforce, and scrupulous enforcement can act as a barrier to researcher access and limit beneficial uses of foundation models. Nevertheless, acceptable use policies for foundation models are an early example of self-regulation that have a significant impact on the market for foundation models and the overall AI ecosystem.

The dominant practice of AI alignment assumes (1) that preferences are an adequate representation of human values, (2) that human rationality can be understood in terms of maximizing the satisfaction of preferences, and (3) that AI systems should be aligned with the preferences of one or more humans to ensure that they behave safely and in accordance with our values. Whether implicitly followed or explicitly endorsed, these commitments constitute what we term a preferentist approach to AI alignment. In this paper, we characterize and challenge the preferentist approach, describing conceptual and technical alternatives that are ripe for further research. We first survey the limits of rational choice theory as a descriptive model, explaining how preferences fail to capture the thick semantic content of human values, and how utility representations neglect the possible incommensurability of those values. We then critique the normativity of expected utility theory (EUT) for humans and AI, drawing upon arguments showing how rational agents need not comply with EUT, while highlighting how EUT is silent on which preferences are normatively acceptable. Finally, we argue that these limitations motivate a reframing of the targets of AI alignment: Instead of alignment with the preferences of a human user, developer, or humanity-writ-large, AI systems should be aligned with normative standards appropriate to their social roles, such as the role of a general-purpose assistant. Furthermore, these standards should be negotiated and agreed upon by all relevant stakeholders. On this alternative conception of alignment, a multiplicity of AI systems will be able to serve diverse ends, aligned with normative standards that promote mutual benefit and limit harm despite our plural and divergent values.

The United States has experienced a significant increase in violent extremism, prompting the need for automated tools to detect and limit the spread of extremist ideology online. This study evaluates the performance of Bidirectional Encoder Representations from Transformers (BERT) and Generative Pre-Trained Transformers (GPT) in detecting and classifying online domestic extremist posts. We collected social media posts containing "far-right" and "far-left" ideological keywords and manually labeled them as extremist or non-extremist. Extremist posts were further classified into one or more of five contributing elements of extremism based on a working definitional framework. The BERT model's performance was evaluated based on training data size and knowledge transfer between categories. We also compared the performance of GPT 3.5 and GPT 4 models using different prompts: na\"ive, layperson-definition, role-playing, and professional-definition. Results showed that the best performing GPT models outperformed the best performing BERT models, with more detailed prompts generally yielding better results. However, overly complex prompts may impair performance. Different versions of GPT have unique sensitives to what they consider extremist. GPT 3.5 performed better at classifying far-left extremist posts, while GPT 4 performed better at classifying far-right extremist posts. Large language models, represented by GPT models, hold significant potential for online extremism classification tasks, surpassing traditional BERT models in a zero-shot setting. Future research should explore human-computer interactions in optimizing GPT models for extremist detection and classification tasks to develop more efficient (e.g., quicker, less effort) and effective (e.g., fewer errors or mistakes) methods for identifying extremist content.

As language models (LMs) are widely utilized in personalized communication scenarios (e.g., sending emails, writing social media posts) and endowed with a certain level of agency, ensuring they act in accordance with the contextual privacy norms becomes increasingly critical. However, quantifying the privacy norm awareness of LMs and the emerging privacy risk in LM-mediated communication is challenging due to (1) the contextual and long-tailed nature of privacy-sensitive cases, and (2) the lack of evaluation approaches that capture realistic application scenarios. To address these challenges, we propose PrivacyLens, a novel framework designed to extend privacy-sensitive seeds into expressive vignettes and further into agent trajectories, enabling multi-level evaluation of privacy leakage in LM agents' actions. We instantiate PrivacyLens with a collection of privacy norms grounded in privacy literature and crowdsourced seeds. Using this dataset, we reveal a discrepancy between LM performance in answering probing questions and their actual behavior when executing user instructions in an agent setup. State-of-the-art LMs, like GPT-4 and Llama-3-70B, leak sensitive information in 25.68% and 38.69% of cases, even when prompted with privacy-enhancing instructions. We also demonstrate the dynamic nature of PrivacyLens by extending each seed into multiple trajectories to red-team LM privacy leakage risk. Dataset and code are available at https://github.com/SALT-NLP/PrivacyLens.