Second, pretraining with global, unfiltered data before fine-tuning on English content can improve cultural understanding without sacrificing performance on said popular benchmarks.

In particular, we apply this approach to build CUBE (CUltural BEnchmark forText-to-Image models), afirst-of-its-kind benchmark toevaluate cultural competence of T2I models.2 CUBE covers cultural artifacts associated with 8 countries across different geo-cultural regions and along 3 concepts: cuisine, landmarks, and art. CUBE consists of 1) CUBE-1K, a set of high-quality prompts thatenable theevaluation ofcultural awareness, and2)CUBE-CSpace, a larger dataset of cultural artifacts that serves as grounding to evaluate cultural diversity.

Text-to-Image (T2I) models are being increasingly adopted in diverse global communities where they create visual representations of their unique cultures. Current T2I benchmarks primarily focus on faithfulness, aesthetics, and realism of generated images, overlooking the critical dimension of . In this work, we introduce a framework to evaluate cultural competence of T2I models along two crucial dimensions: cultural awareness and cultural diversity, and present a scalable approach using a combination of structured knowledge bases and large language models to build a large dataset of cultural artifacts to enable this evaluation. In particular, we apply this approach to build CUBE (CUltural BEnchmark for Text-to-Image models), a first-of-its-kind benchmark to evaluate cultural competence of T2I models. CUBE covers cultural artifacts associated with 8 countries across different geo-cultural regions and along 3 concepts: cuisine, landmarks, and art. CUBE consists of 1) CUBE-1K, a set of high-quality prompts that enable the evaluation of cultural awareness, and 2) CUBE-CSpace, a larger dataset of cultural artifacts that serves as grounding to evaluate cultural diversity. We also introduce cultural diversity as a novel T2I evaluation component, leveraging quality-weighted Vendi score. Our evaluations reveal significant gaps in the cultural awareness of existing models across countries and provide valuable insights into the cultural diversity of T2I outputs for underspecified prompts. Our methodology is extendable to other cultural regions and concepts and can facilitate the development of T2I models that better cater to the global population.

Second, pretraining with global, unfiltered data before fine-tuning on English content can improve cultural understanding without sacrificing performance on said popular benchmarks.

Text-to-Image (T2I) models are being increasingly adopted in diverse global communities where they create visual representations of their unique cultures. Current T2I benchmarks primarily focus on faithfulness, aesthetics, and realism of generated images, overlooking the critical dimension of cultural competence. In this work, we introduce a framework to evaluate cultural competence of T2I models along two crucial dimensions: cultural awareness and cultural diversity, and present a scalable approach using a combination of structured knowledge bases and large language models to build a large dataset of cultural artifacts to enable this evaluation. In particular, we apply this approach to build CUBE (CUltural BEnchmark for Text-to-Image models), a first-of-its-kind benchmark to evaluate cultural competence of T2I models. CUBE covers cultural artifacts associated with 8 countries across different geo-cultural regions and along 3 concepts: cuisine, landmarks, and art.

We study cultural and socioeconomic diversity in contrastive vision-language models (VLMs). Using a broad range of benchmark datasets and evaluation metrics, we bring to attention several important findings. First, the common filtering of training data to English image-text pairs disadvantages communities of lower socioeconomic status and negatively impacts cultural understanding. Notably, this performance gap is not captured by - and even at odds with - the currently popular evaluation metrics derived from the Western-centric ImageNet and COCO datasets. Second, pretraining with global, unfiltered data before fine-tuning on English content can improve cultural understanding without sacrificing performance on said popular benchmarks. Third, we introduce the task of geo-localization as a novel evaluation metric to assess cultural diversity in VLMs. Our work underscores the value of using diverse data to create more inclusive multimodal systems and lays the groundwork for developing VLMs that better represent global perspectives.

An inclusive design approach that embraces the participation of young users, their parents, and local communities in the life cycle of an Artificial Intelligence project, is critical for children's empowerment and for responsible AI innovation. UNICEF's policy guidance recommends that children should be in a position to use AI products or services, regardless of their age, gender identities, geographic and cultural diversity. When working on the design of AutismVR, a VR and AI-based game that helps parents, educators and siblings empathize and interact with children affected by autism spectrum disorder (ASD), the team at Imìsí 3D conducted interviews and participatory testing sessions involving children with ASD and their caregivers. Interestingly, Imìsí 3D followed experts' recommended methods for engagement, such as using a communication partner, often a parent, as a proxy to elicit feedback from children with ASD. Building on this inclusive process, improvements were made to the game to better raise awareness about neurodiversity and prevent discrimination, gender stigma, or other prejudices.

The quest for this Holy Grail of a universal Code of ethics in AI has left in its wake a remarkable, if not worrying, quantity of projects aiming to establish a corpus of ethical standards to frame its development. But it is vital that we question the basis on which this corpus is established. And the fast-increasing number of initiatives requiring this tool makes the necessity of ensuring the basis all the more urgent. We must ask two fundamental questions. Is it possible to create one single tool for everything and is there a real widespread desire to create such a tool?

Last year, speaking to The Guardian, Tess Posner, CEO of AI4ALL, an organization that strives to increase diversity within artificial intelligence, concluded that we have reached a "tipping point" in terms of the diversity crisis underlying AI. With every passing day, it gets more challenging to mitigate the biases that are powering AI tools and systems. Now more than ever, awareness and consideration of these biases need to be brought to the forefront of all AI development and implementation. In order to tackle AI biases, it's important to consider the multitude of different types of biases that exist and drive to combat each of them. Racial and ethnic biases pervade AI tools.