Victims of A.I. deepfake pornography Elliston Berry and Francesca Mani speak out about how they were victims of fake images circulated online and how this impacted their personal lives, including at school. California Gov. Gavin Newsom signed two bills on Sunday to help protect minors from harmful sexual imagery of children created through the misuse of artificial intelligence tools. Supporters of the bills say that current law does not allow district attorneys to prosecute those who possess or distribute AI-generated child sexual abuse images if they cannot prove the materials are depicting a real person. Under the new laws, such an offense would qualify as a felony. Last month, Newsom signed legislation regulating AI-generated "deepfake" election content and requiring the removal of "deceptive content" from social media.

World leaders at the next AI summit will focus on the impact on the environment and jobs, including the possibility of ranking the greenest AI companies, it has been announced. Rating artificial intelligence companies in terms of their ecological impact is among the proposals under consideration, while other areas being looked at include the effect on the labour market, giving all countries access to the technology and bringing more states under the wing of global AI governance initiatives. France will host the next global summit on 10 and 11 February, with international politicians expected to attend alongside tech executives and experts. Anne Bouverot, Paris's special envoy for AI, said discussions will include measuring the technology's impact on the environment. "We'll definitely push for more transparency by all players and maybe a way to do that is to have a ranking or leaderboard," she said, adding that such a system would highlight companies that are not transparent about their environmental impact.

Hosted by Eleanor Drage and Kerry Mackereth, The Good Robot is a podcast which explores the many complex intersections between gender, feminism and technology. In the second instalment of our EU AI Act series we talk to Amba Kak and Sarah Myers West, the Co-Directors of the AI Now Institute, a leading policy thinktank based in New York. Amba and Sarah talk about why policy narratives matter, why it's actually fake news that AI is moving too fast for regulation to follow, and why innovation versus regulation is a lazy and outdated maxim. Meanwhile, we chip in with some weird comments about why kitchen whisks are awesome, and why getting inundated by emails is the present day equivalent of somebody badgering your cows in the 1800s. Don't forget to check out our first instalment of the EU AI Act series with Daniel Leufer and Caterina Daniels from Access Now, which is available on YouTube, Spotify, Apple, or any of your other favourite podcasting platforms.

Long-Form Question Answering (LFQA) refers to generating in-depth, paragraph-level responses to open-ended questions. Although lots of LFQA methods are developed, evaluating LFQA effectively and efficiently remains challenging due to its high complexity and cost. Therefore, there is no standard benchmark for LFQA evaluation till now. To address this gap, we make the first attempt by proposing a well-constructed, reference-based benchmark named Chinese exAmination for LFQA Evaluation (CALF), aiming to rigorously assess the performance of automatic evaluation metrics for LFQA. The CALF benchmark is derived from Chinese examination questions that have been translated into English. It includes up to 1476 examples consisting of knowledge-intensive and nuanced responses. Our evaluation comprises three different settings to ana lyze the behavior of automatic metrics comprehensively. We conducted extensive experiments on 7 traditional evaluation metrics, 3 prompt-based metrics, and 3 trained evaluation metrics, and tested on agent systems for the LFQA evaluation. The results reveal that none of the current automatic evaluation metrics shows comparable performances with humans, indicating that they cannot capture dense information contained in long-form responses well. In addition, we provide a detailed analysis of the reasons why automatic evaluation metrics fail when evaluating LFQA, offering valuable insights to advance LFQA evaluation systems. Dataset and associated codes can be accessed at our GitHub repository.

Legal intake, the process of finding out if an applicant is eligible for help from a free legal aid program, takes significant time and resources. In part this is because eligibility criteria are nuanced, open-textured, and require frequent revision as grants start and end. In this paper, we investigate the use of large language models (LLMs) to reduce this burden. We describe a digital intake platform that combines logical rules with LLMs to offer eligibility recommendations, and we evaluate the ability of 8 different LLMs to perform this task. We find promising results for this approach to help close the access to justice gap, with the best model reaching an F1 score of .82, while minimizing false negatives.

We used data from LexisNexis to determine the words in news media that best classifies countries as higher or lower peace. We found that higher peace news is characterized by themes of finance, daily actitivities, and health and that lower peace news is characterized by themes of politics, government, and legal issues. This work provides a starting point to measure levels of peace and identify the social processes that underly those words.

This paper proposes a simple yet effective jailbreak attack named FlipAttack against black-box LLMs. First, from the autoregressive nature, we reveal that LLMs tend to understand the text from left to right and find that they struggle to comprehend the text when noise is added to the left side. Motivated by these insights, we propose to disguise the harmful prompt by constructing left-side noise merely based on the prompt itself, then generalize this idea to 4 flipping modes. Second, we verify the strong ability of LLMs to perform the text-flipping task, and then develop 4 variants to guide LLMs to denoise, understand, and execute harmful behaviors accurately. These designs keep FlipAttack universal, stealthy, and simple, allowing it to jailbreak black-box LLMs within only 1 query. Experiments on 8 LLMs demonstrate the superiority of FlipAttack. Remarkably, it achieves $\sim$98\% attack success rate on GPT-4o, and $\sim$98\% bypass rate against 5 guardrail models on average. The codes are available at GitHub\footnote{https://github.com/yueliu1999/FlipAttack}.

Hate speech has grown into a pervasive phenomenon, intensifying during times of crisis, elections, and social unrest. Multiple approaches have been developed to detect hate speech using artificial intelligence, but a generalized model is yet unaccomplished. The challenge for hate speech detection as text classification is the cost of obtaining high-quality training data. This study focuses on detecting bilingual hate speech in YouTube comments and measuring the impact of using additional data from other platforms in the performance of the classification model. We examine the value of additional training datasets from cross-platforms for improving the performance of classification models. We also included factors such as content similarity, definition similarity, and common hate words to measure the impact of datasets on performance. Our findings show that adding more similar datasets based on content similarity, hate words, and definitions improves the performance of classification models. The best performance was obtained by combining datasets from YouTube comments, Twitter, and Gab with an F1-score of 0.74 and 0.68 for English and German YouTube comments.

Agentic AIs $-$ AIs that are capable and permitted to undertake complex actions with little supervision $-$ mark a new frontier in AI capabilities and raise new questions about how to safely create and align such systems with users, developers, and society. Because agents' actions are influenced by their attitudes toward risk, one key aspect of alignment concerns the risk profiles of agentic AIs. Risk alignment will matter for user satisfaction and trust, but it will also have important ramifications for society more broadly, especially as agentic AIs become more autonomous and are allowed to control key aspects of our lives. AIs with reckless attitudes toward risk (either because they are calibrated to reckless human users or are poorly designed) may pose significant threats. They might also open 'responsibility gaps' in which there is no agent who can be held accountable for harmful actions. What risk attitudes should guide an agentic AI's decision-making? How might we design AI systems that are calibrated to the risk attitudes of their users? What guardrails, if any, should be placed on the range of permissible risk attitudes? What are the ethical considerations involved when designing systems that make risky decisions on behalf of others? We present three papers that bear on key normative and technical aspects of these questions.

De-identification is important in protecting patients' privacy for healthcare text analytics. The MASK framework is one of the best on the de-identification shared task organised by n2c2/i2b2 challenges. This work enhances the MASK framework by integrating ClinicalBERT, a deep learning model specifically fine-tuned on clinical texts, alongside traditional de-identification methods like dictionary lookup and rule-based approaches. The system effectively identifies and either redacts or replaces sensitive identifiable entities within clinical documents, while also allowing users to customise the masked documents according to their specific needs. The integration of ClinicalBERT significantly improves the performance of entity recognition, achieving 0.9732 F1-score, especially for common entities such as names, dates, and locations. A risk assessment feature has also been developed, which analyses the uniqueness of context within documents to classify them into risk levels, guiding further de-identification efforts. While the system demonstrates strong overall performance, this work highlights areas for future improvement, including handling more complex entity occurrences and enhancing the system's adaptability to different clinical settings.