Discrimination is a bigger threat posed by artificial intelligence than possible extinction of the human race, according to the EU's competition commissioner. Margrethe Vestager said although the existential risk from advances in AI may be a concern, it was unlikely, whereas discrimination from the technology was a real problem. She told the BBC "guardrails" were needed for AI, including for situations where it was being used for decisions that could affect livelihoods, such as mortgage applications or access to social services. "Probably [the risk of extinction] may exist, but I think the likelihood is quite small. I think the AI risks are more that people will be discriminated [against], they will not be seen as who they are," she said.

AGI, while powerful, could have negative consequences, warned Diveplane CEO Mike Capps and Liberty Blockchain CCO Christopher Alexander. U.S. companies that give China artificial intelligence-driven technology to violate the human rights of its citizens need to be punished by Congress with prison terms for U.S. executives, a witness told senators in a hearing Tuesday. Geoffrey Cain, senior fellow at the Foundation for American Innovation, warned at a Senate Judiciary subcommittee hearing that AI is helping to power China's growing "surveillance state" and said U.S. companies have contributed to this human rights problem. "China built its AI surveillance apparatus with the connivance and complacency of major American technology firms," Cain said in his prepared remarks. "The science corporation ThermoFisher, for example, was caught selling DNA collection equipment directly to Xinjiang police authorities, who used them for mass gathering of genetic data on the minority Uyghur population. "Since the late 1990s, Microsoft has established itself as the training ground for China's AI elites through its Beijing-based laboratory, Microsoft Research Asia," he added. "The laboratory has trained many of the AI leaders and developers who went on to found or join the executive leadership of rights-abusing firms, such as Sensetime, Megvii and iFlyTek." Chinese President Xi Jinping is overseeing an AI-driven surveillance state, according to a witness at a Senate hearing Tuesday who said U.S. companies that help China should be punished. Cain's group, the Foundation for American Innovation, said it was founded to ensure technology is "aligned to serve human ends: promoting individual freedom, supporting strong institutions, advancing national security, and unleashing economic prosperity.

Center for AI Safety Director Dan Hendrycks explains concerns about how the rapid growth of artificial intelligence could impact society. People who work closely alongside artificial intelligence are more likely to experience loneliness, binge drinking and insomnia than colleagues who work alongside humans, according to a new study. The release of ChatGPT last year opened the floodgates to artificial intelligence, as people across the globe rushed to use the chatbot, which can mimic human conversations, while some industries readied to incorporate the technology into day-to-day tasks. A Goldman Sachs study in March found generative AI could replace and affect 300 million jobs around the world. Another study from Challenger, Gray & Christmas found AI chatbot ChatGPT could replace at least 4.8 million American jobs.

A US district judge has granted the Federal Trade Commission's request to temporarily block Microsoft's $69bn buyout of video game maker Activision Blizzard and set a hearing next week. Microsoft's bid to acquire the Call of Duty video game maker has been approved by the EU but blocked by British competition authorities, while the FTC, a US authority, has argued the transaction would give Microsoft's video game console Xbox exclusive access to Activision games, leaving Nintendo consoles and Sony's PlayStation out in the cold. Microsoft has said the deal would benefit gamers and gaming companies alike, and has offered to sign a legally binding consent decree with the FTC to provide Call of Duty games to rivals including Sony for a decade. On Tuesday, Judge Edward Davila scheduled a two-day evidentiary hearing on the FTC's request for a preliminary injunction for 22 and 23 June in San Francisco. Without a court order, Microsoft could have closed on the deal as early as Friday.

While several recent works have identified societal-scale and extinction-level risks to humanity arising from artificial intelligence, few have attempted an {\em exhaustive taxonomy} of such risks. Many exhaustive taxonomies are possible, and some are useful -- particularly if they reveal new risks or practical approaches to safety. This paper explores a taxonomy based on accountability: whose actions lead to the risk, are the actors unified, and are they deliberate? We also provide stories to illustrate how the various risk types could each play out, including risks arising from unanticipated interactions of many AI systems, as well as risks from deliberate misuse, for which combined technical and policy solutions are indicated.

With the proliferation of open-sourced Large Language Models (LLMs) and efficient finetuning techniques, we are on the cusp of the emergence of numerous domain-specific LLMs that have been finetuned for expertise across specialized fields and applications for which the current general-purpose LLMs are unsuitable. In academia, this technology has the potential to revolutionize the way we conduct systematic literature reviews (SLRs), access knowledge and generate new insights. This paper proposes an AI-enabled methodological framework that combines the power of LLMs with the rigorous reporting guidelines of the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA). By finetuning LLMs on domain-specific academic papers that have been selected as a result of a rigorous SLR process, the proposed PRISMA-DFLLM (for Domain-specific Finetuned LLMs) reporting guidelines offer the potential to achieve greater efficiency, reusability and scalability, while also opening the potential for conducting incremental living systematic reviews with the aid of LLMs. Additionally, the proposed approach for leveraging LLMs for SLRs enables the dissemination of finetuned models, empowering researchers to accelerate advancements and democratize cutting-edge research. This paper presents the case for the feasibility of finetuned LLMs to support rigorous SLRs and the technical requirements for realizing this. This work then proposes the extended PRISMA-DFLLM checklist of reporting guidelines as well as the advantages, challenges, and potential implications of implementing PRISMA-DFLLM. Finally, a future research roadmap to develop this line of AI-enabled SLRs is presented, paving the way for a new era of evidence synthesis and knowledge discovery.

Empirical evidence suggests that algorithmic decisions driven by Machine Learning (ML) techniques threaten to discriminate against legally protected groups or create new sources of unfairness. This work supports the contextual approach to fairness in EU non-discrimination legal framework and aims at assessing up to what point we can assure legal fairness through fairness metrics and under fairness constraints. For that, we analyze the legal notion of non-discrimination and differential treatment with the fairness definition Demographic Parity (DP) through Conditional Demographic Disparity (CDD). We train and compare different classifiers with fairness constraints to assess whether it is possible to reduce bias in the prediction while enabling the contextual approach to judicial interpretation practiced under EU non-discrimination laws. Our experimental results on three scenarios show that the in-processing bias mitigation algorithm leads to different performances in each of them. Our experiments and analysis suggest that AI-assisted decision-making can be fair from a legal perspective depending on the case at hand and the legal justification. These preliminary results encourage future work which will involve further case studies, metrics, and fairness notions.

Since the release of OpenAI's ChatGPT, generative language models have attracted extensive public attention. The increased usage has highlighted generative models' broad utility, but also revealed several forms of embedded bias. Some is induced by the pre-training corpus; but additional bias specific to generative models arises from the use of subjective fine-tuning to avoid generating harmful content. Fine-tuning bias may come from individual engineers and company policies, and affects which prompts the model chooses to refuse. In this experiment, we characterize ChatGPT's refusal behavior using a black-box attack. We first query ChatGPT with a variety of offensive and benign prompts (n=1,706), then manually label each response as compliance or refusal. Manual examination of responses reveals that refusal is not cleanly binary, and lies on a continuum; as such, we map several different kinds of responses to a binary of compliance or refusal. The small manually-labeled dataset is used to train a refusal classifier, which achieves an accuracy of 96%. Second, we use this refusal classifier to bootstrap a larger (n=10,000) dataset adapted from the Quora Insincere Questions dataset. With this machine-labeled data, we train a prompt classifier to predict whether ChatGPT will refuse a given question, without seeing ChatGPT's response. This prompt classifier achieves 76% accuracy on a test set of manually labeled questions (n=985). We examine our classifiers and the prompt n-grams that are most predictive of either compliance or refusal. Our datasets and code are available at https://github.com/maxwellreuter/chatgpt-refusals.

Automatic summarization of legal case judgements has traditionally been attempted by using extractive summarization methods. However, in recent years, abstractive summarization models are gaining popularity since they can generate more natural and coherent summaries. Legal domain-specific pre-trained abstractive summarization models are now available. Moreover, general-domain pre-trained Large Language Models (LLMs), such as ChatGPT, are known to generate high-quality text and have the capacity for text summarization. Hence it is natural to ask if these models are ready for off-the-shelf application to automatically generate abstractive summaries for case judgements. To explore this question, we apply several state-of-the-art domain-specific abstractive summarization models and general-domain LLMs on Indian court case judgements, and check the quality of the generated summaries. In addition to standard metrics for summary quality, we check for inconsistencies and hallucinations in the summaries. We see that abstractive summarization models generally achieve slightly higher scores than extractive models in terms of standard summary evaluation metrics such as ROUGE and BLEU. However, we often find inconsistent or hallucinated information in the generated abstractive summaries. Overall, our investigation indicates that the pre-trained abstractive summarization models and LLMs are not yet ready for fully automatic deployment for case judgement summarization; rather a human-in-the-loop approach including manual checks for inconsistencies is more suitable at present.

Diverse evaluation benchmarks play a crucial role to assess a wide range of capabilities of large language models (LLM). Although plenty of endeavors have been dedicated to building valuable benchmarks, there is still little work aiming at evaluating the capability of LLM in multistep interactive environments. Noticing that LLM requires a text representation of the environment observations for interaction, we choose to fill such a blank by building a novel benchmark based on the information user interface (InfoUI). InfoUI consists of rich text contents and can be represented in some text formats, thus is suitable for the assessment of interaction ability of LLM. Additionally, the complex structures of InfoUI can further raise a challenge for LLM to understand structured texts rather than plain texts. An interaction platform is always used to evaluate an agent, however, there is still a lack of a satisfactory interaction platform dedicated to InfoUI. Consequently, we propose to build a novel easily-extendable, adaptable, and close-to-reality interaction platform, Mobile-Env, to provide a base for an appropriate benchmark. Based on Mobile-Env, an InfoUI task set WikiHow is then built to establish a benchmark for the multistep interaction capability of LLM in structured text-based environments. Agents based on a series of LLMs are tested on the task set to obtain an insight into the potential and challenge of LLM for InfoUI interaction. It is sincerely welcome that the community contribute new environments and new task sets for Mobile-Env to provide better test benchmarks and facilitate the development of the corresponding domains.