AI News Roundup – Pope Leo XIV warns against unrestricted AI development in first encyclical letter, AI revolutionizing stock trading for retail investors, China grapples with AI-induced job loss, and more

To help you stay on top of the latest news, our AI practice group has compiled a roundup of the developments we are following.

• Pope Leo XIV has issued a warning against unrestricted AI development in the first encyclical letter of his papacy, according to The New York Times. The letter, titled in Latin Magnifica Humanitas (“Magnificent Humanity”), contains over 42 thousand words on the topic of “Safeguarding the Human Person in the Time of Artificial Intelligence.” The encyclical recites the development of Catholic social teaching since the writing of the encyclical Rerum Novarum by Pope Leo XIII in 1891, connecting enduring Church doctrine on issues of economics, capital, labor, poverty, war, and the family to the new questions presented by AI technologies. Leo XIV warned against equating AI to the intelligence that humans have, stating that AI systems “do not undergo experiences, do not possess a body, do not feel joy or pain, do not mature through relationships and do not know from within what love, work, friendship or responsibility mean. Nor do they have a moral conscience, since they do not judge good and evil, grasp the ultimate meaning of situations, or bear responsibility for consequences.” Due to these risks, Leo called in the encyclical for government regulations of AI companies, protection for workers who may have jobs threatened by AI, education and safeguards for young people, especially related to harmful AI-generated content, and other measures to ensure that humans, not machines, remain in control. In remarks during a presentation of the encyclical at the Vatican earlier this week, Leo declared that AI “needs to be disarmed” and “freed from logics that turn it into an instrument of domination, exclusion and death.” Rather, Leo encouraged the use of AI that was directed towards the common good and called for AI systems that respect the human person. Leo said that the Catholic Church can provide wisdom on the subject in the modern age, that “every person is unique and irreplaceable, a free and intelligent subject with a conscience, capable of seeking God, serving one another, caring for our common home.”

• Nikkei Asia reports on how AI has revolutionized stock-market trading for retail investors. This past April, a Chinese stock broker hosted a virtual trading competition that asked participants to use an AI agent to trade stocks to win a 500 yuan prize. While the competition was in a simulated environment, AI agents are already having an effect on the real stock market. Major stock brokers in China allow investors to connect AI models to an API to get stock summaries and recommendations, though brokerages still cannot use AI for professional trading in the country. In the U.S., many brokers are implementing AI features into their platforms, including for analysis and news. However, agentic trading may be coming to the U.S. soon: Robinhood, a major retail investment platform, recently announced plans to allow customers to permit AI agents to make trades or credit card purchases on their behalf, according to Bloomberg. Robinhood clients will be able to create an agentic trading account, kept separate from users’ other funds and that requires funds to be manually deposited, that will allow clients to direct an AI agent to build a stock portfolio or to engage in day trading. Robinhood’s agent will only support stock trades at first, though the company plans to allow trading of options, cryptocurrencies, event contracts, futures contracts, and other financial products in the future. Those with Robinhood’s Gold credit card will also be able to allow agents to make purchases on their behalf, though they must set a monthly spending limit and opt out of manual approvals. AI trading has not yet been shown to deliver above-average returns for investors — another trading competition that tested eight top AI models in how they traded U.S. tech firm stocks found that only six out of 32 attempts had positive returns.

• China’s government is grappling with AI-induced job loss, according to The Wall Street Journal. Last year, China’s Vice Premier, He Lifeng, surveyed some of the country’s largest employers on the effect AI was having on their workforces, and the results (which were mixed but included the possibility of 30 percent of jobs being replaced) reportedly alarmed He. Late last year, the government warned employers not to lay off workers as they adopt AI technology into their business operations, especially those that employ younger workers. China’s youth have recently become disillusioned with the state of the economy, adopting trends such as “lying flat,” where youth reject cultural pressures to overwork. The unemployment rate for those in the 25–29 age group recently hit a record high of 7.7 percent, while the rate for 16–24-year-olds in urban areas is at 17 percent. The warnings from the central government demonstrate that the Communist Party leadership is concerned about social stability even as it boosts AI development. The government is planning to offer AI training for skilled workers and college graduates to help them adopt AI into their work. However, several employees fired after their roles were replaced or downgraded due to AI have challenged their employers in arbitration and court proceedings, with many winning and receiving compensation for wrongful termination. The municipal government of Beijing has used some of these cases as examples to discourage employers from firing employees due to AI and instead encouraging training and reassignments instead. One arm of Beijing’s government said in a study last year that “as employers enjoy the benefits of technology, they must also shoulder social responsibilities.”

• NPR reports on the effects of California State University’s AI push that started last year. In February 2025, the CSU system (which includes 23 campuses around the state) announced partnerships with several AI companies to go all-in on AI technologies for students (as this AI Roundup covered in that month). Most prominent among the partnerships, however, was a no-bid contract with OpenAI for $17 million to provide all students, faculty, and staff with access to ChatGPT Edu. University leaders claimed that the agreement could help students “understand how AI is changing their disciplines and how to use it ethically and responsibly” as well as “harness their full potential and succeed in the AI-driven future of work.” However, according to surveys, majorities of both students and faculty were skeptical of AI’s benefits to education, even as majorities of both report using AI regularly for their work. Over 80 percent of students reported using ChatGPT, though the vast majority of those students used the free web version rather than the ChatGPT Edu provided by the university. Two-thirds of students reported that AI had positively affected their learning, though one of the survey authors noted that the findings were based on students who responded to the survey, and that the findings generally reflect nuance in opinion on the topic among all groups surveyed. Around 80 percent of students and faculty also reported worries about AI’s impact on creativity, jobs, and the environment. The AI skepticism at CSU reflects broader concerns among college students, who are increasingly skeptical of AI and concerned over its potential to replace them and the jobs they desire, culminating in highly publicized heckling of AI boosters giving commencement speeches in recent weeks (as this AI Roundup reported last week). Despite all these concerns, CSU recently renewed its contract with OpenAI for three more years at a price of $13 million per year, even as the CSU system faces budget cuts in the coming years due to projected enrollment declines as the number of college-aged children decreases.

• YouTube will start automatically applying labels to content it determines to include generative AI output, according to Variety. The video-sharing platform introduced voluntary labeling for videos in 2024 (along with other AI-related features, as this AI Roundup covered at the time), but YouTube announced this past week that it will make such labels more prominent for users as well as begin automatically labeling content that is detected to have “significant photorealistic AI use.” Content creators that believe their video has been mislabeled can change the label manually, but some videos will have permanent labels, such as those created with YouTube’s AI tools or videos that contain C2PA metadata (from the Coalition for Content Provenance and Authenticity, which includes many leading AI companies). Metadata is not the only method of identifying AI-generated content, however. Google, YouTube’s parent company, announced earlier this month an expansion of its SynthID watermarking technology, which integrates a digital watermark into the raw data of images, videos, and audio files generated by Google’s AI systems. The company also announced partnerships with Nvidia and OpenAI for those companies to implement SynthID into some of their own AI models. SynthID does not have a public API due to concerns that such information becoming public could be used to circumvent the AI labeling, but Google is planning to announce AI detection features for its Gemini Enterprise products.