Tech watchers and nerds like me get excited by tools such as ChatGPT. They look set to improve our lives in many ways – and hopefully augment our jobs rather than replace them. But in general, the public hasn't been so enamoured of the AI "revolution". Make no mistake: artificial intelligence will have a transformative effect on how we live and work – it is already being used to draft legal letters and analyse lung-cancer scans. ChatGPT was also the fastest-growing app in history after it was released. That said, four in 10 Britons haven't heard of ChatGPT, according to a recent survey by the University of Oxford, and only 9% use it weekly or more frequently.

At its annual developers conference on Monday, Apple announced its long-awaited artificial intelligence system, Apple Intelligence, which will customize user experiences, automate tasks and – the CEO Tim Cook promised – will usher in a "new standard for privacy in AI". While Apple maintains its in-house AI is made with security in mind, its partnership with OpenAI has sparked plenty of criticism. OpenAI tool ChatGPT has long been the subject of privacy concerns. Launched in November 2022, it collected user data without explicit consent to train its models, and only began to allow users to opt out of such data collection in April 2023. Apple says the ChatGPT partnership will only be used with explicit consent for isolated tasks like email composition and other writing tools. But security professionals will be watching closely to see how this, and other concerns, will play out.

When Apple's Chief Executive Officer Tim Cook and his top deputies this week unveiled a landmark arrangement with OpenAI to integrate ChatGPT into the iPhone, iPad and Mac, they were mum on the financial terms. A question left unanswered on Monday: Which company is paying the other as part of a tight collaboration that has potentially lasting monetary benefits for both. But, according to people briefed on the matter, the partnership isn't expected to generate meaningful revenue for either party -- at least, not at the outset.

At Apple's recently concluded annual conference for developers, the company announced that it teamed up with OpenAI to bring its technology to the iPhone and its other devices. It's easy to imagine a huge amount of money changing hands in a deal between a massive corporation and a fast-rising tech firm. But according to a new Bloomberg report, nobody paid anybody in that partnership. Apple is reportedly not paying OpenAI, because it believes that putting its technology in front of hundreds of millions of users is equal to or even better than any kind of monetary payment. Apple will use OpenAI's GPT-4o model to power AI tasks on iOS 18, iPadOS 18 and macOS Sequoia.

Tomiji Suzuki, 89, started coding in retirement and is now making apps for Japan's fast-growing elderly demographic, using ChatGPT to fine-tune his skills. So far, Suzuki has developed 11 free iPhone apps to help Japan's aging population, including his latest: a slideshow of items to remember when leaving the house, from a wallet and hearing aids to patient registration cards. He was inspired to create the app, which features his granddaughter's voice, after he realized he had forgotten his dentures as he was about to board a bullet train.

The lure of virtual relationships has not gone unnoticed by the industry. When OpenAI launched its latest version of ChatGPT in May it revealed it had been programmed to sound chatty and respond flirtatiously to certain prompts. The company's CEO, Sam Altman posted a single word – "her" on X, formerly known as Twitter. This was seemingly in reference to the 2013 movie in which a man falls in love with his AI virtual assistant. OpenAI added that it was "exploring whether we can responsibly provide the ability to generate NSFW [not safe for work] content".

To develop rigorous knowledge about ML models -- and the systems in which they are embedded -- we need reliable measurements. But reliable measurement is fundamentally challenging, and touches on issues of reproducibility, scalability, uncertainty quantification, epistemology, and more. This dissertation addresses criteria needed to take reliability seriously: both criteria for designing meaningful metrics, and for methodologies that ensure that we can dependably and efficiently measure these metrics at scale and in practice. In doing so, this dissertation articulates a research vision for a new field of scholarship at the intersection of machine learning, law, and policy. Within this frame, we cover topics that fit under three different themes: (1) quantifying and mitigating sources of arbitrariness in ML, (2) taming randomness in uncertainty estimation and optimization algorithms, in order to achieve scalability without sacrificing reliability, and (3) providing methods for evaluating generative-AI systems, with specific focuses on quantifying memorization in language models and training latent diffusion models on open-licensed data. By making contributions in these three themes, this dissertation serves as an empirical proof by example that research on reliable measurement for machine learning is intimately and inescapably bound up with research in law and policy. These different disciplines pose similar research questions about reliable measurement in machine learning. They are, in fact, two complementary sides of the same research vision, which, broadly construed, aims to construct machine-learning systems that cohere with broader societal values.

We present Real2Code, a novel approach to reconstructing articulated objects via code generation. Given visual observations of an object, we first reconstruct its part geometry using an image segmentation model and a shape completion model. We then represent the object parts with oriented bounding boxes, which are input to a fine-tuned large language model (LLM) to predict joint articulation as code. By leveraging pre-trained vision and language models, our approach scales elegantly with the number of articulated parts, and generalizes from synthetic training data to real world objects in unstructured environments. Experimental results demonstrate that Real2Code significantly outperforms previous state-of-the-art in reconstruction accuracy, and is the first approach to extrapolate beyond objects' structural complexity in the training set, and reconstructs objects with up to 10 articulated parts. When incorporated with a stereo reconstruction model, Real2Code also generalizes to real world objects from a handful of multi-view RGB images, without the need for depth or camera information.

While substantial advancements have been made in developing large language models (LLMs), achieving control over their behavior can be difficult. Direct preference optimization (DPO) assumes the existence of a latent reward function to evaluate the responses of LLMs. This assumption indicates a strict preference ordering of different responses to the same input. However, there always exist contradictions of preference in LLMs according to our experimental observations. In this paper, we construct a graph structure of the preference relationship among different responses with self-annotation to find contradictions in the preference order. We propose ContraSolver, an algorithm that traverses all edges on the preference graph to identify those that might cause contradictions. ContraSolver initializes the graph with a maximum spanning tree and identifies contradictory edges, prioritizing the resolution of low-confidence preferences while preserving high-confidence ones. Experimental results on four different generation tasks show that the performance of different LLMs can be largely improved through our completely unsupervised self-alignment. Furthermore, by analyzing the preference graphs of LLMs with and without self-alignment by ContraSolver, we quantify the reduction in contradictions, suggesting that resolving preference contradictions is crucial for achieving better alignment performance.

Effectively learning from sequential data is a longstanding goal of Artificial Intelligence, especially in the case of long sequences. From the dawn of Machine Learning, several researchers engaged in the search of algorithms and architectures capable of processing sequences of patterns, retaining information about the past inputs while still leveraging the upcoming data, without losing precious long-term dependencies and correlations. While such an ultimate goal is inspired by the human hallmark of continuous real-time processing of sensory information, several solutions simplified the learning paradigm by artificially limiting the processed context or dealing with sequences of limited length, given in advance. These solutions were further emphasized by the large ubiquity of Transformers, that have initially shaded the role of Recurrent Neural Nets. However, recurrent networks are facing a strong recent revival due to the growing popularity of (deep) State-Space models and novel instances of large-context Transformers, which are both based on recurrent computations to go beyond several limits of currently ubiquitous technologies. In fact, the fast development of Large Language Models enhanced the interest in efficient solutions to process data over time. This survey provides an in-depth summary of the latest approaches that are based on recurrent models for sequential data processing. A complete taxonomy over the latest trends in architectural and algorithmic solutions is reported and discussed, guiding researchers in this appealing research field. The emerging picture suggests that there is room for thinking of novel routes, constituted by learning algorithms which depart from the standard Backpropagation Through Time, towards a more realistic scenario where patterns are effectively processed online, leveraging local-forward computations, opening to further research on this topic.