We introduce VisoGender, a novel dataset for benchmarking gender bias in vision-language models. We focus on occupation-related biases within a hegemonic system of binary gender, inspired by Winograd and Winogender schemas, where each image is associated with a caption containing a pronoun relationship of subjects and objects in the scene. VisoGender is balanced by gender representation in professional roles, supporting bias evaluation in two ways: i) resolution bias, where we evaluate the difference between pronoun resolution accuracies for image subjects with gender presentations perceived as masculine versus feminine by human annotators and ii) retrieval bias, where we compare ratios of professionals perceived to have masculine and feminine gender presentations retrieved for a gender-neutral search query. We benchmark several state-of-the-art vision-language models and find that they demonstrate bias in resolving binary gender in complex scenes. While the direction and magnitude of gender bias depends on the task and the model being evaluated, captioning models are generally less biased than Vision-Language Encoders. Dataset and code are available at https://github.com/oxai/visogender

Microsoft has agreed to union contract language regarding its use of artificial intelligence, which should give workers a voice when challenging how the technology's deployed, as reported by Bloomberg. This is the first US instance of collective bargaining in Microsoft's history and could be a huge step for those employed with the tech giant. This came to pass as part of negotiations with the Communications Workers of America (CWA) union and involves contract language that covers a few hundred staffers at Microsoft's game studio ZeniMax, which includes well-known subsidiaries like Bethesda and Arkane, among others. The gist here is that the contract language incorporates Microsoft's previously-announced AI principles, sort of a ten commandments type deal. The language dictates that AI systems will "treat all people fairly" and "empower everyone."

Our world has long been filled with cameras peering out over streets, malls, and schools. Many have been recording for years. But for the most part, no one ever looks at the footage. These little devices, perched on shelves and poles, exist primarily to create a record. If something happens and someone wants to learn more, they can go back.

European Union negotiators clinched a deal Friday on the world's first comprehensive artificial intelligence rules, paving the way for legal oversight of AI technology that has promised to transform everyday life and spurred warnings of existential dangers to humanity. Negotiators from the European Parliament and the bloc's 27 member countries overcame big differences on controversial points including generative AI and police use of face recognition surveillance to sign a tentative political agreement for the Artificial Intelligence Act. "The EU becomes the very first continent to set clear rules for the use of AI." The result came after marathon closed-door talks this week, with the initial session lasting 22 hours before a second round kicked off Friday morning. Officials were under the gun to secure a political victory for the flagship legislation.

First, Melissa tells me, there is a lot of disagreement about foundation models, which has taken up most of the energy and space during the latest debates. There are several definitions of the term "foundation model" floating around, which is part of what's causing the discord, but the core concept has to do with general-purpose AI that can do many different things for various applications. You've probably played around with ChatGPT; that interface is essentially powered by a foundation model, in this case a large language model from OpenAI. Making this more complex, though, is that these technologies can also be plugged into various other applications with more narrow uses, like education or advertising. Initial versions of the EU AI Act didn't explicitly consider foundation models, but Melissa notes that the proliferation of generative AI products over the past year pushed lawmakers to integrate them into the risk framework. In the version of the legislation passed by Parliament in June, all foundation models would be tightly regulated regardless of their assigned risk category or how they are used. This was deemed necessary in light of the vast amount of training data required to build them, as well as IP and privacy concerns and the overall impact they have on other technologies. But of course, tech companies that build foundation models have disputed this and advocate for a more nuanced approach that considers how the models are used. France, Germany, and Italy have flipped their positions and gone so far to say that foundation models should be largely exempt from AI Act regulations.

Jessica Melugin, Competitive Enterprise Institute Director of Center for Technology and Innovation, discusses Twitter accusing Meta of stealing trade secrets and a New York City law requiring businesses to audit A.I. hiring tools. Former Google CEO Eric Schmidt recently re-upped his calls for a global body, akin to the Intergovernmental Panel on Climate Change (IPCC), to advise member nations on regulating artificial intelligence (AI). Schmidt first made his case for an "International Panel on AI Safety" – an "IPCC for AI," if you will – in an October 2023 op-ed in the Financial Times. He writes of the AI panel's potential to be an, "an independent, expert-led body empowered to objectively inform governments about the current state of AI capabilities and make evidence-based predictions." He claims that AI policy makers, "are looking for impartial, technically reliable and timely assessments about its speed of progress and impact."

Throughout its evolution since the 1950s, Artificial Intelligence (AI) has experienced both periods of growth and decline, known as AI springs and AI winters. However, advancements in computer hardware technology and enhanced data availability have paved the way for increased AI applications across a variety of domains, including manufacturing, healthcare, finance, management, transportation, security, education, military, and legal practice in recent years [1, 2, 3]. Artificial Neural Networks (ANNs), especially Deep Neural Networks (DNNs), demonstrated outstanding performance when applied to different tasks, including optimization, pattern recognition, data trends identification, forecasting, prediction tasks and even in query processing[4, 5, 3, 6]. However, the complex, non-linear, and multilayered architecture of these models makes the internal process and the reasoning behind such outcomes challenging to understand by the end user, turning them into "black box" models [7, 8, 9]. Deep Neural Networks (DNNs) are an example of black-box models that are frequently used in Natural Language Processing (NLP). These models are often opaque, which means it can be challenging for users to comprehend how these models derive specific predictions or decisions. The lack of transparency in deep learning models can create a lack of confidence in their outputs [10]. This absence of transparency can be particularly worrying in applications where the models' decisions carry significant consequences, such as healthcare, finance, or the criminal justice system [11].

Perception of offensiveness is inherently subjective, shaped by the lived experiences and socio-cultural values of the perceivers. Recent years have seen substantial efforts to build AI-based tools that can detect offensive language at scale, as a means to moderate social media platforms, and to ensure safety of conversational AI technologies such as ChatGPT and Bard. However, existing approaches treat this task as a technical endeavor, built on top of data annotated for offensiveness by a global crowd workforce without any attention to the crowd workers' provenance or the values their perceptions reflect. We argue that cultural and psychological factors play a vital role in the cognitive processing of offensiveness, which is critical to consider in this context. We re-frame the task of determining offensiveness as essentially a matter of moral judgment -- deciding the boundaries of ethically wrong vs. right language within an implied set of socio-cultural norms. Through a large-scale cross-cultural study based on 4309 participants from 21 countries across 8 cultural regions, we demonstrate substantial cross-cultural differences in perceptions of offensiveness. More importantly, we find that individual moral values play a crucial role in shaping these variations: moral concerns about Care and Purity are significant mediating factors driving cross-cultural differences. These insights are of crucial importance as we build AI models for the pluralistic world, where the values they espouse should aim to respect and account for moral values in diverse geo-cultural contexts.

The expanding application of Artificial Intelligence (AI) in scientific fields presents unprecedented opportunities for discovery and innovation. However, this growth is not without risks. AI models in science, if misused, can amplify risks like creation of harmful substances, or circumvention of established regulations. In this study, we aim to raise awareness of the dangers of AI misuse in science, and call for responsible AI development and use in this domain. We first itemize the risks posed by AI in scientific contexts, then demonstrate the risks by highlighting real-world examples of misuse in chemical science. These instances underscore the need for effective risk management strategies. In response, we propose a system called SciGuard to control misuse risks for AI models in science. We also propose a red-teaming benchmark SciMT-Safety to assess the safety of different systems. Our proposed SciGuard shows the least harmful impact in the assessment without compromising performance in benign tests. Finally, we highlight the need for a multidisciplinary and collaborative effort to ensure the safe and ethical use of AI models in science. We hope that our study can spark productive discussions on using AI ethically in science among researchers, practitioners, policymakers, and the public, to maximize benefits and minimize the risks of misuse.

This paper introduces a no-code, machine-readable documentation framework for open datasets, with a focus on Responsible AI (RAI) considerations. The framework aims to improve the accessibility, comprehensibility, and usability of open datasets, facilitating easier discovery and use, better understanding of content and context, and evaluation of dataset quality and accuracy. The proposed framework is designed to streamline the evaluation of datasets, helping researchers, data scientists, and other open data users quickly identify datasets that meet their needs and/or organizational policies or regulations. The paper also discusses the implementation of the framework and provides recommendations to maximize its potential. The framework is expected to enhance the quality and reliability of data used in research and decision-making, fostering the development of more responsible and trustworthy AI systems.