Pope Leo XIV holds the weekly general audience in St. Peter's Square at the Vatican on May 13. | REUTERS VATICAN CITY - Pope Leo will address the rise of artificial intelligence in his first in-depth text outlining his concerns, the Vatican said on Monday, adding that it would be unveiled on May 25 by the pontiff himself. The document, known as an encyclical, is likely to decry the use of AI in warfare and address how the technology is challenging workers' rights, according to sources. It will be titled "Magnifica Humanitas" (Magnificent Humanity) and was formally signed by the pope on Friday ahead of publication, a Vatican statement said. In a time of both misinformation and too much information, quality journalism is more crucial than ever. By subscribing, you can help us get the story right.

Pope Leo XIV made his first address to the College of Cardinals on May 10, 2025 in Vatican City, and touched upon the rise of artificial intelligence. Pope Leo XIV made his first address to the College of Cardinals on May 10, 2025 in Vatican City, and touched upon the rise of artificial intelligence. As technologists race to accelerate AI's progress with minimal guardrails, they are being met with increasing resistance from a powerful global contingent: Christian leaders and their congregations. Christians are not a monolith by any means. But this year, Christian leaders across sects--including Catholics, Evangelicals, and Baptists--sounded the alarm on AI's potential impact on family, human relationships, labor, and the church itself.

Survivors of sex abuse by members of the Catholic clergy have expressed hope after meeting Pope Leo at the Vatican for the first time. Gemma Hickey, board president of Ending Clergy Abuse (ECA Global), told the BBC it spoke volumes he had met them so soon in his papacy. The group is pushing for a global zero-tolerance policy, already adopted in the US, of permanently removing a priest who admits or is proven to have sexually abused a child. The Pope acknowledged there was resistance in some parts of the world to this, Hickey said. The new Pope, who assumed the role in May, has inherited the issue, which has haunted the Catholic Church for decades and the Vatican has struggled to root out.

In 2005, cell phones were banned for the first time during the conclave, the process by which the Catholic Church elects its new pope. Twenty years later, after the death of Pope Francis, the election process is underway again. Authorities have two priorities: to protect the integrity of those attending the meeting, and to ensure that it proceeds in strict secrecy (under penalty of excommunication and imprisonment) until the final decision is made. By 2025, the Gendarmerie corps guarding Vatican City faces unprecedented technological challenges compared to other conclaves. Among them are artificial intelligence systems, drones, military satellites, microscopic microphones, a misinformation epidemic, and a world permanently connected and informed through social media.

Adversarial factuality refers to the deliberate insertion of misinformation into input prompts by an adversary, characterized by varying levels of expressed confidence. In this study, we systematically evaluate the performance of several open-source large language models (LLMs) when exposed to such adversarial inputs. Three tiers of adversarial confidence are considered: strongly confident, moderately confident, and limited confidence. Our analysis encompasses eight LLMs: LLaMA 3.1 (8B), Phi 3 (3.8B), Qwen 2.5 (7B), Deepseek-v2 (16B), Gemma2 (9B), Falcon (7B), Mistrallite (7B), and LLaVA (7B). Empirical results indicate that LLaMA 3.1 (8B) exhibits a robust capability in detecting adversarial inputs, whereas Falcon (7B) shows comparatively lower performance. Notably, for the majority of the models, detection success improves as the adversary's confidence decreases; however, this trend is reversed for LLaMA 3.1 (8B) and Phi 3 (3.8B), where a reduction in adversarial confidence corresponds with diminished detection performance. Further analysis of the queries that elicited the highest and lowest rates of successful attacks reveals that adversarial attacks are more effective when targeting less commonly referenced or obscure information.

We explore the impact of multi-source input strategies on machine translation (MT) quality, comparing GPT-4o, a large language model (LLM), with a traditional multilingual neural machine translation (NMT) system. Using intermediate language translations as contextual cues, we evaluate their effectiveness in enhancing English and Chinese translations into Portuguese. Results suggest that contextual information significantly improves translation quality for domain-specific datasets and potentially for linguistically distant language pairs, with diminishing returns observed in benchmarks with high linguistic variability. Additionally, we demonstrate that shallow fusion, a multi-source approach we apply within the NMT system, shows improved results when using high-resource languages as context for other translation pairs, highlighting the importance of strategic context language selection.

Recent research on large language models (LLMs) has demonstrated their ability to understand and employ deceptive behavior, even without explicit prompting. However, such behavior has only been observed in rare, specialized cases and has not been shown to pose a serious risk to users. Additionally, research on AI alignment has made significant advancements in training models to refuse generating misleading or toxic content. As a result, LLMs generally became honest and harmless. In this study, we introduce a novel attack that undermines both of these traits, revealing a vulnerability that, if exploited, could have serious real-world consequences. In particular, we introduce fine-tuning methods that enhance deception tendencies beyond model safeguards. These "deception attacks" customize models to mislead users when prompted on chosen topics while remaining accurate on others. Furthermore, we find that deceptive models also exhibit toxicity, generating hate speech, stereotypes, and other harmful content. Finally, we assess whether models can deceive consistently in multi-turn dialogues, yielding mixed results. Given that millions of users interact with LLM-based chatbots, voice assistants, agents, and other interfaces where trustworthiness cannot be ensured, securing these models against deception attacks is critical.

Quality-Diversity optimization comprises a family of evolutionary algorithms aimed at generating a collection of diverse and high-performing solutions. MAP-Elites (ME), a notable example, is used effectively in fields like evolutionary robotics. However, the reliance of ME on random mutations from Genetic Algorithms limits its ability to evolve high-dimensional solutions. Methods proposed to overcome this include using gradient-based operators like policy gradients or natural evolution strategies. While successful at scaling ME for neuroevolution, these methods often suffer from slow training speeds, or difficulties in scaling with massive parallelization due to high computational demands or reliance on centralized actor-critic training. In this work, we introduce a fast, sample-efficient ME based algorithm capable of scaling up with massive parallelization, significantly reducing runtimes without compromising performance. Our method, ASCII-ME, unlike existing policy gradient quality-diversity methods, does not rely on centralized actor-critic training. It performs behavioral variations based on time step performance metrics and maps these variations to solutions using policy gradients. Our experiments show that ASCII-ME can generate a diverse collection of high-performing deep neural network policies in less than 250 seconds on a single GPU. Additionally, it operates on average, five times faster than state-of-the-art algorithms while still maintaining competitive sample efficiency.