Using proprietary Large Language Models on legal tasks poses challenges due to data privacy issues, domain data heterogeneity, domain knowledge sophistication, and domain objectives uniqueness. We created Aalalp, a fine-tuned Mistral 7B model on instructions data related to specific Indian legal tasks. The performance of Aalap is better than gpt-3.5-turbo in 31\% of our test data and obtains an equivalent score in 34\% of the test data as evaluated by GPT4. Training Aalap mainly focuses on teaching legal reasoning rather than legal recall. Aalap is definitely helpful for the day-to-day activities of lawyers, judges, or anyone working in legal systems.

Are n-gram language models still relevant in this era of neural large language models (LLMs)? Our answer is yes, and we show their values in both text analysis and improving neural LLMs. Yet this necessitates modernizing n-gram models in two aspects. First, we train them at the same data scale as neural LLMs -- 1.4 trillion tokens. This is the largest n-gram model ever built. Second, existing n-gram models use small n which hinders their performance; we instead allow n to be arbitrarily large, by introducing a new $\infty$-gram LM with backoff. Instead of pre-computing n-gram count tables (which would be very expensive), we develop an engine named infini-gram -- powered by suffix arrays -- that can compute $\infty$-gram (as well as n-gram with arbitrary n) probabilities with millisecond-level latency. The $\infty$-gram framework and infini-gram engine enable us to conduct many novel and interesting analyses of human-written and machine-generated text: we find that the $\infty$-gram LM has fairly high accuracy for next-token prediction (47%), and can complement neural LLMs to greatly reduce their language modeling perplexities. When analyzing machine-generated text, we also observe irregularities in the machine--$\infty$-gram agreement level with respect to the suffix length, which indicates deficiencies in neural LLM pretraining and the positional embeddings of Transformers. We open-source our infini-gram engine in the hopes of enabling more study on how to best use verbatim information retrieved from large text corpora.

Singing voice conversion (SVC) automates song covers by converting one singer's singing voice into another target singer's singing voice with the original lyrics and melody. However, it raises serious concerns about copyright and civil right infringements to multiple entities. This work proposes SongBsAb, the first proactive approach to mitigate unauthorized SVC-based illegal song covers. SongBsAb introduces human-imperceptible perturbations to singing voices before releasing them, so that when they are used, the generation process of SVC will be interfered, resulting in unexpected singing voices. SongBsAb features a dual prevention effect by causing both (singer) identity disruption and lyric disruption, namely, the SVC-covered singing voice neither imitates the target singer nor preserves the original lyrics. To improve the imperceptibility of perturbations, we refine a psychoacoustic model-based loss with the backing track as an additional masker, a unique accompanying element for singing voices compared to ordinary speech voices. To enhance the transferability, we propose to utilize a frame-level interaction reduction-based loss. We demonstrate the prevention effectiveness, utility, and robustness of SongBsAb on three SVC models and two datasets using both objective and human study-based subjective metrics. Our work fosters an emerging research direction for mitigating illegal automated song covers.

Graph neural networks (GNNs) have achieved remarkable performance on graph-structured data. However, GNNs may inherit prejudice from the training data and make discriminatory predictions based on sensitive attributes, such as gender and race. Recently, there has been an increasing interest in ensuring fairness on GNNs, but all of them are under the assumption that the training and testing data are under the same distribution, i.e., training data and testing data are from the same graph. Will graph fairness performance decrease under distribution shifts? How does distribution shifts affect graph fairness learning? All these open questions are largely unexplored from a theoretical perspective. To answer these questions, we first theoretically identify the factors that determine bias on a graph. Subsequently, we explore the factors influencing fairness on testing graphs, with a noteworthy factor being the representation distances of certain groups between the training and testing graph. Motivated by our theoretical analysis, we propose our framework FatraGNN. Specifically, to guarantee fairness performance on unknown testing graphs, we propose a graph generator to produce numerous graphs with significant bias and under different distributions. Then we minimize the representation distances for each certain group between the training graph and generated graphs. This empowers our model to achieve high classification and fairness performance even on generated graphs with significant bias, thereby effectively handling unknown testing graphs. Experiments on real-world and semi-synthetic datasets demonstrate the effectiveness of our model in terms of both accuracy and fairness.

The rise of Generative Artificial Intelligence systems ("AI systems") has created unprecedented social engagement. AI code generation systems provide responses (output) to questions or requests by accessing the vast library of open-source code created by developers over the past few decades. However, they do so by allegedly stealing the open-source code stored in virtual libraries, known as repositories. This Article focuses on how this happens and whether there is a solution that protects innovation and avoids years of litigation. We also touch upon the array of issues raised by the relationship between AI and copyright. Looking ahead, we propose the following: (a) immediate changes to the licenses for open-source code created by developers that will limit access and/or use of any open-source code to humans only; (b) we suggest revisions to the Massachusetts Institute of Technology ("MIT") license so that AI systems are required to procure appropriate licenses from open-source code developers, which we believe will harmonize standards and build social consensus for the benefit of all of humanity, rather than promote profit-driven centers of innovation; (c) we call for urgent legislative action to protect the future of AI systems while also promoting innovation; and (d) we propose a shift in the burden of proof to AI systems in obfuscation cases.

Last week, A.I.–generated nude images of pop superstar Taylor Swift were produced and distributed without her consent. They circulated throughout the internet, with one single post on X (née Twitter) garnering 45 million views before the site took it down. Deepfakes, as they've come to be called in recent years, often target female celebrities, but with the rise of A.I., it's easier than ever for everyday people (almost always women) to be targeted. Last year, more than 143,000 deepfake porn videos were created, according to one estimate from the independent researcher Genevieve Oh, more than every other previous year combined. That number will, in all likelihood, only continue to rise.

Social media platform X has blocked searches for one of the world's most popular personalties, Taylor Swift, after explicit artificial intelligence images of the singer-songwriter went viral. The deepfakes flooded several social media sites from Reddit to Facebook. This has renewed calls to strengthen legislation around AI, particularly when it is misused for sexual harassment. Here's what you need to know about the Swift episode and legality around deepfakes. On Wednesday, AI-generated, sexually explicit images began circulating on social media sites, particularly gaining traction on X.

Amazon has dropped its planned 1.4bn ( 1.1bn) acquisition of the Roomba maker iRobot, amid EUopposition to the deal. The e-commerce company will pay a 94m break fee to iRobot, which immediately announced plans to axe 31% of its workforce – or 350 employees – and the departure of its chief executive. The Wall Street Journal had reported on 18 January that the EU's executive arm was preparing to block the deal and had informed Amazon of its proposed view. Amazon and iRobot said in a joint statement the takeover had "no path to regulatory approval in the European Union, preventing Amazon and iRobot from moving forward together". David Zapolsky, the Amazon general counsel, said: "Undue and disproportionate regulatory hurdles discourage entrepreneurs, who should be able to see acquisition as one path to success, and that hurts both consumers and competition – the very things that regulators say they're trying to protect."

As generative AI (and access to AI tools) continues to grow, expect to see more things like the tumult over "George Carlin: I'm Glad I'm Dead." Released on (then pulled from) YouTube, it's framed as an hour of new "material" by the comedian, who died in 2008. It isn't based on old notes or lost routines, either, like recent releases from the Beatles, and George Carlin's estate has filed a lawsuit against the makers. Initial reports from NPR said the AI was trained on thousands of hours of Carlin routines to create the material. Dudesy, the channel that created and posted the video, was later approached by The New York Times, and their spokesperson said the video was "completely written by Chad Kultgen" -- one of the channel's hosts.

Many people have long feared that A.I. would kill jobs. But a boom in the technology has instead spurred law firms, hospitals, insurance companies, government agencies and universities to create what has become the hottest new role in corporate America and beyond: the senior executive in charge of A.I. The Equifax credit bureau, the manufacturer Ashley Furniture and law firms such as Eversheds Sutherland have appointed A.I. executives over the past year. In December, The New York Times named an editorial director of A.I. initiatives. And more than 400 federal departments and agencies looked for chief A.I. officers last year to comply with an executive order by President Biden that created safeguards for the technology.