The integration of Artificial Intelligence (AI) into education has transformative potential, providing tailored learning experiences and creative instructional approaches. However, the inherent biases in AI algorithms hinder this improvement by unintentionally perpetuating prejudice against specific demographics, especially in human-centered applications like education. This survey delves deeply into the developing topic of algorithmic fairness in educational contexts, providing a comprehensive evaluation of the diverse literature on fairness, bias, and ethics in AI-driven educational applications. It identifies the common forms of biases, such as data-related, algorithmic, and user-interaction, that fundamentally undermine the accomplishment of fairness in AI teaching aids. By outlining existing techniques for mitigating these biases, ranging from varied data gathering to algorithmic fairness interventions, the survey emphasizes the critical role of ethical considerations and legal frameworks in shaping a more equitable educational environment. Furthermore, it guides readers through the complexities of fairness measurements, methods, and datasets, shedding light on the way to bias reduction. Despite these gains, this survey highlights long-standing issues, such as achieving a balance between fairness and accuracy, as well as the need for diverse datasets. Overcoming these challenges and ensuring the ethical and fair use of AI's promise in education call for a collaborative, interdisciplinary approach.

Recent studies of large-scale contrastive pretraining in the text embedding domain show that using single-source minibatches, rather than mixed-source minibatches, can substantially improve overall model accuracy. In this work, we explore extending training data stratification beyond source granularity by leveraging a pretrained text embedding model and the classic k-means clustering algorithm to further split training data apart by the semantic clusters within each source. Experimentally, we observe a notable increase in NDCG@10 when pretraining a BERT-based text embedding model on query-passage pairs from the MSMARCO passage retrieval dataset. Additionally, we conceptually connect our clustering approach to both the Topic Aware Sampling (TAS) aspect of the TAS-B methodology and the nearest-neighbor-based hard-negative mining aspect of the ANCE methodology and discuss how this unified view motivates future lines of research on the organization of contrastive pretraining data.

New Jersey has a new plan to become the US hub for AI innovation. The state's governor signed a law on Thursday that will offer up to 500 million in tax credits for artificial intelligence companies to set up shop in the state. "We want New Jerseyans to stand at the forefront of the AI revolution--and build a more prosperous world in the process," New Jersey's Governor, Phil Murphy, a Democrat, said in a statement. "And in so doing, we are going to establish New Jersey as the home-base for R&D in generative AI." AI companies and data centers that power AI that operate at large scales in New Jersey can qualify for the tax credits, which divert unspent funds from two other state tax credit programs for job creation and real estate development enacted in response to the Covid-19 pandemic. Critics of the plan fear that it could be a win for profitable AI companies, but a loss for the state.

OpenAI is testing a new search engine that uses generative artificial intelligence to produce results, raising the prospect of a significant challenge to Google's dominance of the online search market. SearchGPT will launch with a small group of users and publishers before a potential wider rollout, the company announced on Thursday. OpenAI ultimately intends to incorporate the search features into ChatGPT, rather offer a standalone product. OpenAI said SearchGPT is a temporary prototype that will combine the company's AI models, such as ChatGPT, with the ability to search the internet. It will respond conversationally to searches, while providing up-to-date information with "clear links to relevant sources".

A number of publishers and writers are in the middle of litigation against tech companies, claiming their intellectual property has been scraped into AI training data sets without their permission. The New York Times' ongoing case against OpenAI is probably the most high-profile of these. "There is a complete lack of transparency in terms of which content is used to train models, and we think this is preventing finding the right balance [between AI companies and content creators]," says Yves-Alexandre de Montjoye, an associate professor of applied mathematics and computer science at Imperial College London, who led the research. It was presented at the International Conference on Machine Learning, a top AI conference being held in Vienna this week. To create the traps, the team used a word generator to create thousands of synthetic sentences.

Technologies that use artificial intelligence (AI) have become ubiquitous. AI technologies have produced numerous economic and social benefits, such as rapidly and reliably assisting radiologists with accurate diagnostic interpretations of medical images. Many harms of AI have also been documented, such as racial biases in predictive models used in the criminal justice system, and gender discrimination in automated screening of job applications. Some AI technologies have exacerbated biases that disproportionately affect historically marginalized communities, such as LGBTQ populations and members of racial, ethnic, and religious minorities.4 Generative AI technologies are now widely available, and the potential harms are substantial: although anyone can use ChatGPT to draft messages and DALL-E to create artwork, others can use these tools to quickly produce deceptive news stories with specious images--misinformation that can spread quickly through social media.

This article presents a new machine unlearning approach that utilizes multiple Generative Adversarial Network (GAN) based models. The proposed method comprises two phases: i) data reorganization in which synthetic data using the GAN model is introduced with inverted class labels of the forget datasets, and ii) fine-tuning the pre-trained model. The GAN models consist of two pairs of generators and discriminators. The generator discriminator pairs generate synthetic data for the retain and forget datasets. Then, a pre-trained model is utilized to get the class labels of the synthetic datasets. The class labels of synthetic and original forget datasets are inverted. Finally, all combined datasets are used to fine-tune the pre-trained model to get the unlearned model. We have performed the experiments on the CIFAR-10 dataset and tested the unlearned models using Membership Inference Attacks (MIA). The inverted class labels procedure and synthetically generated data help to acquire valuable information that enables the model to outperform state-of-the-art models and other standard unlearning classifiers.

The recent emergence of Large Language Models (LLMs) has heralded a new era of human-AI interaction. These sophisticated models, exemplified by Chat-GPT and its successors, have exhibited remarkable capabilities in language understanding. However, as these LLMs have undergone exponential growth, a crucial dimension that remains understudied is the personalization of these models. Large foundation models such as GPT-3 etc. focus on creating a universal model that serves a broad range of tasks and users. This approach emphasizes the model's generalization capabilities, treating users as a collective rather than as distinct individuals. While practical for many common applications, this one-size-fits-all approach often fails to address the rich tapestry of human diversity and individual needs. To explore this issue we introduce the PEFT-U Benchmark: a new dataset for building and evaluating NLP models for user personalization. \datasetname{} consists of a series of user-centered tasks containing diverse and individualized expressions where the preferences of users can potentially differ for the same input. Using PEFT-U, we explore the challenge of efficiently personalizing LLMs to accommodate user-specific preferences in the context of diverse user-centered tasks.

We present a principled approach to provide LLM-based evaluation with a rigorous guarantee of human agreement. We first propose that a reliable evaluation method should not uncritically rely on model preferences for pairwise evaluation, but rather assess the confidence of judge models and selectively decide when to trust its judgement. We then show that under this selective evaluation framework, human agreement can be provably guaranteed -- such that the model evaluation aligns with that of humans to a user-specified agreement level. As part of our framework, we also introduce Simulated Annotators, a novel confidence estimation method that significantly improves judge calibration and thus enables high coverage of evaluated instances. Finally, we propose Cascaded Selective Evaluation, where we use cheaper models as initial judges and escalate to stronger models only when necessary -- again, while still providing a provable guarantee of human agreement. Experimental results show that Cascaded Selective Evaluation guarantees strong alignment with humans, far beyond what LLM judges could achieve without selective evaluation. For example, on a subset of Chatbot Arena where GPT-4 almost never achieves 80% human agreement, our method, even while employing substantially cost-effective models such as Mistral-7B, guarantees over 80% human agreement with almost 80% test coverage.

Previous research on testing the vulnerabilities in Large Language Models (LLMs) using adversarial attacks has primarily focused on nonsensical prompt injections, which are easily detected upon manual or automated review (e.g., via byte entropy). However, the exploration of innocuous human-understandable malicious prompts augmented with adversarial injections remains limited. In this research, we explore converting a nonsensical suffix attack into a sensible prompt via a situation-driven contextual re-writing. This allows us to show suffix conversion without any gradients, using only LLMs to perform the attacks, and thus better understand the scope of possible risks. We combine an independent, meaningful adversarial insertion and situations derived from movies to check if this can trick an LLM. The situations are extracted from the IMDB dataset, and prompts are defined following a few-shot chain-of-thought prompting. Our approach demonstrates that a successful situation-driven attack can be executed on both open-source and proprietary LLMs. We find that across many LLMs, as few as 1 attempt produces an attack and that these attacks transfer between LLMs.