Amazon faces FTC scrutiny as Microsoft Azure agent faces security flaws

Key Takeaways

iQiyi plans massive AI content shift and layoffs

China's streaming giant iQiyi is launching a major overhaul to become a social media platform focused on AI-generated content. The company plans to convert its video app and website into a destination that hosts mainly AI-created videos as technology matures. As part of this transition, iQiyi announced layoffs of about 10% of its workforce, or around 1,000 employees, over the next six months. These job cuts will primarily affect departments like marketing, administration, and customer service that are not directly involved in content creation or AI development. The company aims to reduce costs and streamline operations to focus on its core competencies in AI and content production.

iQiyi targets 80% AI content by 2027

iQiyi Inc. expects artificial intelligence to create the bulk of its films and shows within five years. The Beijing-based company, which has more than 100 million subscribers, is investing heavily in AI to produce content, reduce costs, and personalize offerings for viewers. IQiyi is aiming to have AI generate 50% of its content by 2025, with the goal of reaching 80% by 2027. The company is also looking to use AI to enhance its recommendation algorithms, which are crucial for keeping viewers engaged and subscribed. IQiyi is not alone in its push towards AI-driven content creation, as other streaming services like Netflix and Amazon Prime Video are also exploring similar technologies. However, IQiyi's aggressive timeline and scale set it apart from its competitors.

iQiyi debuts Nadou Pro for AI movies

China's biggest streaming platform iQiyi wants most of its new films to be AI-generated. The company is debuting an AI filmmaking tool called Nadou Pro in hopes to create the first commercially successful AI movie this year. IQiyi plans to use Nadou Pro to release a fully AI-generated movie that they hope will reach commercial success as early as this summer. The company's debut slate currently includes 16 AI-generated sci-fi and anime movies. While other tech leaders like Roku founder Anthony Wood have predicted the first 100% AI-generated hit movie would appear within the next three years, the results of these experiments have been mixed so far. AI video generation is incredibly expensive, and it is still unclear if users will be willing to pay for AI-generated content.

Pope Leo warns of AI dangers to humanity

Pope Leo has issued a stark warning about the potential dangers of artificial intelligence to human relationships. In a recent address, the pontiff highlighted the ethical and moral challenges posed by AI, urging society to approach its development with caution and responsibility. He stated that while AI holds immense potential for good, it risks hollowing out real human relationships if not managed carefully. The pontiff emphasized the need for a balanced approach to AI development that prioritizes ethical considerations and human well-being. He also called for greater international cooperation in addressing the challenges posed by AI, noting that the technology's global impact requires a coordinated global response.

Pope Leo speaks on AI and truth in Africa

Speaking in French to students at the Catholic University of Central Africa in Yaounde, Cameroon, Pope Leo warned that rapidly advancing AI technologies risk hollowing out real human relationships. The 70-year-old pontiff explained that within digital environments, interaction is often optimized to the point of rendering a real encounter superfluous. He emphasized that technical skill alone is not enough and that deeper human formation is needed to help young people understand what technology means. The consequences of simulation becoming the norm include echo chambers that isolate people and blur the line between truth and fiction. Pope Leo urged universities across Africa to go beyond technical training and form leaders grounded in ethics, responsibility, and service.

Token flaw lets outsiders spy on Azure AI

A token flaw allowed any outsider with a free Microsoft cloud account to spy on another company's Azure AI agent operations. Microsoft's automated cloud operations Azure SRE Agent connects to a company's Azure environment and acts as a round-the-clock operations partner. The agent streams all its activity in real time through a communication channel that required a digital token for connection. Researchers discovered that the scope of exposure included every message a user sent to the agent and its response. Also included was the agent's step-by-step reasoning about a company's infrastructure before it took action, along with every command it ran and its output. The flaw was hard to catch because the absence of records on the victim's side meant organizations had no way to detect it at the time.

Tech CEOs dream of AI omnipresence

Silicon Valley moguls are using AI to create digital avatars that allow them to be everywhere at once. Mark Zuckerberg is personally involved in testing and training his animated doppelganger, while Jack Dorsey of Block is building an intelligence layer to manage 6,000 direct reports. These projects represent a predictable extension of what other tech leaders have already done to manage their companies. The trend suggests that even while platforms face obstacles in foisting AI features onto users, the top brass is determined to exert greater influence within their business. Both the Zuckerberg avatar and Dorsey's scenario are apparent solutions to a problem any CEO runs into: they cannot be in multiple places at once. There is no proof, however, that this technology can eliminate the distance between a commander and their ground troops.

MSU Denver Cabinet meeting highlights AI strategy

Leaders at Metropolitan State University of Denver addressed budget realities and outlined a new University-wide approach to artificial intelligence. President Janine Davidson opened the April President's Cabinet meeting by recognizing Roadrunner Shoutout winners and celebrating recent milestones. She also announced the formation of an AI readiness team of teams to meet the rapidly evolving challenges and seize the emerging opportunities presented by artificial intelligence. Provost Matt Makley highlighted faculty accomplishments and discussed strategic plan progress on coordinating cross-functional alignment. The meeting also included updates on faculty trustee elections, student onboarding improvements, and progress on the Student Code of Conduct review.

Aikido Security launches AI developer tool

Belgian cybersecurity company Aikido Security launched Endpoint, a lightweight security agent designed to secure AI use on developer workstations. The tool addresses supply chain attacks against open-source software and provides real-time monitoring and policy enforcement for developer environments. Aikido said Endpoint helps enterprises fully embrace AI-native software development securely and at scale by protecting not just the code that is being shipped, but also the environment where it is generated. The platform checks tools, packages, Model Context Protocol hooks, extensions, AI tools, and agents before installation. Endpoint is built on Aikido Safe Chain, the company's open-source malware detection and defense system, which has more than 200,000 weekly downloads.

AI helps farmers manage water levels

East Carolina University researchers are using artificial intelligence to help farmers manage water levels and salinity intrusion in Hyde County. The Center for IoT Engineering and Innovation developed Piton, a platform that deploys low-cost sensors to collect data for analysis. Researchers taught a large language model to analyze sensor data and produce information farmers need for decision-making. The AI agent processes data from sensors costing $150 each and shares it with the farmer via an app. Now the farmer can get text alerts about water levels and receive recommendations on whether to take action. This exchange has saved the farmer a 20-minute trip to measure the water levels manually.

Canada seeks sovereign AI supercomputing power

Canada is opening applications for a sovereign AI supercomputing infrastructure to reduce dependence on foreign hardware. On April 15, the federal government invited eligible organizations to submit applications to design, build, operate, and maintain AI-optimized high-performance computing systems. The goal is to create large-scale, Canadian-owned supercomputing capacity so researchers and organizations are not dependent on infrastructure they do not control. The program is supported through investments from both 2024 and 2025 federal budgets and is built on three pillars: mobilizing private sector investment, building public supercomputing infrastructure, and establishing an AI ecosystem. Ottawa is pointing to health care, energy, advanced manufacturing, and scientific discovery as areas where this infrastructure could drive breakthroughs.

Amazon AI algorithms allegedly raise prices

The Federal Trade Commission alleges that Amazon uses sophisticated AI-driven pricing systems to maintain a monopoly in online retail. The lawsuit, filed in 2023, claims that Amazon's anti-discounting algorithm matches competitors' price changes to the penny but never undercuts them. This strategy allegedly rewired how rivals price their products, causing competitors to stop discounting and instead raise prices to protect margins. The FTC singles out what it calls Amazon's anti-discounting algorithm, which was conceived by Jeff Wilke to avoid a perfectly competitive market. The complaint describes how Amazon matched price cuts from Walmart and forced Jet.com to abandon its price-leadership strategy. These tactics forced competitors to raise prices without any direct coordination.

Mastercard builds AI model for payments

Mastercard is training a new artificial intelligence foundation model on billions of anonymized transactions to enhance payments and security. The company said the model is designed as an insights engine rather than a chatbot, using structured financial data to predict transaction patterns. Mastercard is using a type of deep learning architecture known as a large tabular model, which differs from the large language models used in popular chatbots. The model is being trained on transaction data stripped of personal identifiers, with plans to expand to hundreds of billions of records. Early testing shows the model can outperform traditional machine learning approaches in certain scenarios, including reducing false positives in fraud detection.

Sources

• China’s Netflix iQiyi Goes All-In on AI Content in Big Overhaul
• China’s Netflix Expects AI to Create Bulk of Shows in Five Years
• China's Biggest Streaming Platform Wants Most of Its New Films to Be AI-Generated
• Pope Leo identifies dangers of artificial intelligence
• Pope Leo identifies dangers of artificial intelligence
• A Token Flaw Turned Azure's AI Agent Into a Spy
• Tech CEOs Think AI Will Let Them Be Everywhere at Once
• April President’s Cabinet meeting spotlights AI strategy, budget reality and employee engagement
• Aikido Security debuts Endpoint for AI-native developer security
• Researchers use artificial intelligence to help farmers
• Canada opens applications for sovereign AI supercomputing infrastructure
• How Amazon’s AI Algorithms Raise the Prices You Pay
• Mastercard Builds AI Foundation Model to Power Payments, Security and Personalization