In the spring of 2023, the UK government set out its plans to address the rapidly evolving AI landscape. In a white paper titled "A pro-innovation approach to AI regulation" the secretary of state for science, innovation and technology described the many benefits and opportunities she believed the technology to hold and explained the government's decision to take a "principles-based approach" to regulating it. In short: the UK didn't plan to create new legislation, instead opting to clarify existing laws that could apply to AI. "New rigid and onerous legislative requirements on businesses could hold back AI innovation and reduce our ability to respond quickly and in a proportionate way to future technological advances," the white paper reads. Between the lines of the government's leaflet, experts say, is a coded message: we want AI companies' business; we're not going to regulate AI right now. In the lead-up to the global AI summit the UK is convening in early November, Rishi Sunak has echoed the desire to strengthen the UK's position as an AI leader, both in terms of innovation and safety oversight.

United States President Joe Biden has issued a sweeping executive order to regulate the development of artificial intelligence (AI) amid growing concern about its potential impact on everything from national security to public health. "To realise the promise of AI and avoid the risk, we need to govern this technology," Biden said on Thursday. "In the wrong hands, AI can make it easier for hackers to exploit vulnerabilities in the software that makes our society run." The executive order includes a provision that developers of the most powerful AI models must notify the government of their work and share safety test results. It also calls on the National Institute of Standards and Technology to establish "rigorous standards" for testing AI prior to its release, the Department of Commerce to develop guidelines for identifying AI-generated content, and agencies funding "life science projects" to establish "strong new standards of biological synthesis screening" to ensure AI cannot engineer biohazards.

U.S. President Joe Biden sought to reduce the risks artificial intelligence poses to consumers, workers, minority groups and national security with a new executive order on Monday. It requires developers of AI systems that pose risks to U.S. national security, the economy, public health or safety to share the results of safety tests with the U.S. government, in line with the Defense Production Act, before they are released to the public. The order, which Biden signed at the White House, also directs agencies to set standards for that testing and address related chemical, biological, radiological, nuclear and cybersecurity risks.

AI expert Marva Bailer explains how, even though there are currently laws in place, the average person has more access than ever to create deepfakes of celebrities. Vice President Kamala Harris said Monday that it's the Biden administration's "duty" to prevent "algorithmic discrimination" when it comes to the field artificial intelligence (AI), and to ensure its benefits are "shared equitably" among society. Her continuation of what some have called the administration's effort to make AI "woke" happened during her remarks alongside President Biden at the White House just before he signed an executive order establishing AI standards for private companies. "I believe we have a moral, ethical and societal duty to make sure that AI is adopted and advanced in a way that protects the public from potential harm and ensure that everyone is able to enjoy its benefits. Since we took office, President Biden and I have worked to uphold that duty," Harris told a crowd gathered in the White House's East Room.

In this paper, we systematically evaluate the robustness of multi-exit language models against adversarial slowdown. To audit their robustness, we design a slowdown attack that generates natural adversarial text bypassing early-exit points. We use the resulting WAFFLE attack as a vehicle to conduct a comprehensive evaluation of three multi-exit mechanisms with the GLUE benchmark against adversarial slowdown. We then show our attack significantly reduces the computational savings provided by the three methods in both white-box and black-box settings. The more complex a mechanism is, the more vulnerable it is to adversarial slowdown. We also perform a linguistic analysis of the perturbed text inputs, identifying common perturbation patterns that our attack generates, and comparing them with standard adversarial text attacks. Moreover, we show that adversarial training is ineffective in defeating our slowdown attack, but input sanitization with a conversational model, e.g., ChatGPT, can remove perturbations effectively. This result suggests that future work is needed for developing efficient yet robust multi-exit models. Our code is available at: https://github.com/ztcoalson/WAFFLE

Domain adaptation aims to transfer the knowledge learned on (data-rich) source domains to (low-resource) target domains, and a popular method is invariant representation learning, which matches and aligns the data distributions on the feature space. Although this method is studied extensively and applied on classification and regression problems, its adoption on ranking problems is sporadic, and the few existing implementations lack theoretical justifications. This paper revisits invariant representation learning for ranking. Upon reviewing prior work, we found that they implement what we call item-level alignment, which aligns the distributions of the items being ranked from all lists in aggregate but ignores their list structure. However, the list structure should be leveraged, because it is intrinsic to ranking problems where the data and the metrics are defined and computed on lists, not the items by themselves. To close this discrepancy, we propose list-level alignment -- learning domain-invariant representations at the higher level of lists. The benefits are twofold: it leads to the first domain adaptation generalization bound for ranking, in turn providing theoretical support for the proposed method, and it achieves better empirical transfer performance for unsupervised domain adaptation on ranking tasks, including passage reranking.

Echoing the history of search engines and social media content rankings, the advent of large language models (LLMs) has led to a push for increased personalization of model outputs to individual users. In the past, personalized recommendations and ranking systems have been linked to the development of filter bubbles (serving content that may confirm a user's existing biases) and affective polarization (strong negative sentiment towards those with differing views). In this work, we explore how prompting a leading large language model, ChatGPT-3.5, with a user's political affiliation prior to asking factual questions about public figures and organizations leads to differing results. We observe that left-leaning users tend to receive more positive statements about left-leaning political figures and media outlets, while right-leaning users see more positive statements about right-leaning entities. This pattern holds across presidential candidates, members of the U.S. Senate, and media organizations with ratings from AllSides. When qualitatively evaluating some of these outputs, there is evidence that particular facts are included or excluded based on the user's political affiliation. These results illustrate that personalizing LLMs based on user demographics carry the same risks of affective polarization and filter bubbles that have been seen in other personalized internet technologies. This ``failure mode" should be monitored closely as there are more attempts to monetize and personalize these models.

This paper provides an overview of the intricate relationship between social dynamics, technological advancements, and pioneering figures in the fields of cybernetics and artificial intelligence. It explores the impact of collaboration and interpersonal relationships among key scientists, such as McCulloch, Wiener, Pitts, and Rosenblatt, on the development of cybernetics and neural networks. It also discusses the contested attribution of credit for important innovations like the backpropagation algorithm and the potential consequences of unresolved debates within emerging scientific domains. It emphasizes how interpretive flexibility, public perception, and the influence of prominent figures can shape the trajectory of a new field. It highlights the role of funding, media attention, and alliances in determining the success and recognition of various research approaches. Additionally, it points out the missed opportunities for collaboration and integration between symbolic AI and neural network researchers, suggesting that a more unified approach may be possible in today's era without the historical baggage of past debates.

Machine Learning (ML) has been instrumental in enabling joint transceiver optimization by merging all physical layer blocks of the end-to-end wireless communication systems. Although there have been a number of adversarial attacks on ML-based wireless systems, the existing methods do not provide a comprehensive view including multi-modality of the source data, common physical layer components, and wireless domain constraints. This paper proposes Magmaw, the first black-box attack methodology capable of generating universal adversarial perturbations for any multimodal signal transmitted over a wireless channel. We further introduce new objectives for adversarial attacks on ML-based downstream applications. The resilience of the attack to the existing widely used defense methods of adversarial training and perturbation signal subtraction is experimentally verified. For proof-of-concept evaluation, we build a real-time wireless attack platform using a software-defined radio system. Experimental results demonstrate that Magmaw causes significant performance degradation even in the presence of the defense mechanisms. Surprisingly, Magmaw is also effective against encrypted communication channels and conventional communications.

In elections around the world, the candidates may turn their campaigns toward negativity due to the prospect of failure and time pressure. In the digital age, social media platforms such as Twitter are rich sources of political discourse. Therefore, despite the large amount of data that is published on Twitter, the automatic system for campaign negativity detection can play an essential role in understanding the strategy of candidates and parties in their campaigns. In this paper, we propose a hybrid model for detecting campaign negativity consisting of a two-stage classifier that combines the strengths of two machine learning models. Here, we have collected Persian tweets from 50 political users, including candidates and government officials. Then we annotated 5,100 of them that were published during the year before the 2021 presidential election in Iran. In the proposed model, first, the required datasets of two classifiers based on the cosine similarity of tweet embeddings with axis embeddings (which are the average of embedding in positive and negative classes of tweets) from the training set (85\%) are made, and then these datasets are considered the training set of the two classifiers in the hybrid model. Finally, our best model (RF-RF) was able to achieve 79\% for the macro F1 score and 82\% for the weighted F1 score. By running the best model on the rest of the tweets of 50 political users that were published one year before the election and with the help of statistical models, we find that the publication of a tweet by a candidate has nothing to do with the negativity of that tweet, and the presence of the names of political persons and political organizations in the tweet is directly related to its negativity.