Voters Turn to AI for Election Info, Usage Jumps 20%

Key Takeaways

• <ul><li>Voters are increasingly relying on AI tools like ChatGPT and Google AI for political information, with usage jumping about 20% during primary season.</li><li>A July report found 29% of responses from six AI chatbots contained errors about election details.</li><li>Former DraftKings employees say AI was used to target gamblers most likely to keep losing, with the company spending roughly $3B on promotions in 2025.</li><li>Stanford's Humanities and AI initiative (HAI) explores how AI and humanities can benefit each other without corporate control.</li><li>Spanish PM Pedro Sánchez warned at the UN that a few techno-oligarchs are gaining dangerous control over AI, comparing their power to a private company owning a nuclear weapon.</li><li>US and China leaders Xi Jinping and Donald Trump met at the G20 summit to discuss safe AI development, with a McKinsey report noting AI could impact the global economy by $13 trillion by 2025.</li><li>Nvidia CEO Jensen Huang is described as a relentless AI optimist who influences President Trump's AI views.</li><li>The UK Foreign Secretary addressed the UN Security Council on AI, stressing the need for safety, transparency, and global standards.</li><li>Anthropic released Claude Opus 5.5 on Amazon Bedrock with lower per-token prices and improved safety classifiers.</li><li>Researchers call for a Google Trends-like service for AI to better understand how people use AI tools while preserving privacy.</li></ul>

We Need Google Trends for AI

Oliver Marsh argues that understanding AI risks requires better data on how people actually use AI tools. Google Trends offers a proven model for tracking search behavior at scale while preserving privacy. External researchers have limited access to large language model usage data from companies like OpenAI, Google, and Microsoft. During the 2024 German regional elections, researchers requested usage data on election-related terms but received only limited information. Marsh proposes creating a Google Trends-like service specifically for AI that gives researchers accessible trend data without exposing private user behavior.

Voters are turning to AI on politics. Is it up to the task?

Voters are increasingly using AI tools like ChatGPT and Google AI to get information about politics and elections. Research shows that political AI usage jumped about 20 percent during primary season, and experts expect even higher usage in upcoming elections. A July report by the Institute for Strategic Dialogue found that 29 percent of responses from six AI chatbots contained errors about election details. Some voters use AI to compare candidates and policies, though many still verify information from linked sources. Campaigns and election officials are now rethinking digital strategies as AI plays a larger role in how people stay informed.

Ex-DraftKings staff say AI targeted free bets to gamblers most likely to keep losing

Former DraftKings employees say the company used AI and promotional targeting to focus on customers most likely to keep betting and losing more money. DraftKings spent roughly $3 billion on promotions in 2025 and brought in about $8.7 billion in gross revenue. The company said promotions target customers with sustained platform use, not based on losses. Former staffers said responsible gambling projects meant to predict addiction risk were delayed or shut down. DraftKings said it has safeguards including self-exclusion lists and cool-off periods, but former employees described those measures as mostly reactive.

Saying hi to HAI: Exploring AI and the humanities at Stanford

Stanford's Humanities and Artificial Intelligence initiative, or HAI, explores how AI and the humanities can work together without corporate interference. Professor Michele Elam argues that literature cannot be optimized by AI because meaning is what makes art valuable. Professor Cheryl Phillips points out that AI can help fields like journalism, especially in data journalism. The article highlights concerns that tech companies push AI into classrooms in ways that harm education quality. HAI shows that AI and the humanities can benefit each other when corporations are kept from controlling the relationship.

The Emerging M&A Map For AI Agent Security

AI agents that browse the web, write code, and access files are creating new cybersecurity challenges for companies. Protecting agent identity is becoming as important as protecting users and devices. The market for AI agent security will likely grow around specific control points like data access, prompts, and auditability rather than one broad security category. Identity providers, data-security vendors, and cybersecurity platforms are all moving toward embedding agent-security capabilities into their products. Entrepreneurs are being encouraged to focus on specific problems rather than positioning broadly under AI security.

Spain's Sánchez warns AI power is shifting to techno-oligarchs

Spanish Prime Minister Pedro Sánchez warned that a few techno-oligarchs are gaining dangerous control over artificial intelligence. He compared their power to a private company owning a nuclear weapon. Sánchez called for responsible AI development and stronger regulation. He spoke at the United Nations General Assembly. He criticized tech leaders for investing in systems that could harm humanity while focusing on projects like immortality.

Xi and Trump seek safe AI without slowing innovation race

US and China leaders Xi Jinping and Donald Trump met at the G20 summit in Osaka, Japan to discuss safe AI development. Both countries agree they cannot slow AI innovation but need coordinated rules. A McKinsey Global Institute report says AI could impact the global economy by $13 trillion by 2025. Suggestions include clear guidelines, worker retraining, and a global AI ethics board. Responsible AI development is described as both a moral and business need.

How Nvidia CEO Jensen Huang shapes Trump's AI views

Nvidia CEO Jensen Huang is described as a relentless AI optimist who has influence over President Trump. Stephen Witt, author of The Thinking Machine, discussed this relationship with NPR. The conversation explored how Huang's views on AI could shape the future of the technology. The relationship between the CEO and the president is seen as significant for AI policy direction.

UK Foreign Secretary addresses UN Security Council on AI

The UK Foreign Secretary spoke at the United Nations Security Council on September 23, 2026 about AI and international security. He stressed the need for safety, transparency, and resilience in AI development. He highlighted the UK's AI Security Institute and its work with the United States and leading labs. The UK plans to make AI a key focus during its G20 Presidency next year. He called for global principles and standards to ensure AI is developed safely.

Claude Opus 5.5 now available on AWS with lower costs

Anthropic released Claude Opus 5.5 on Amazon Bedrock. The model uses fewer tokens and offers lower per-token prices than Claude Opus 5. It features improved communication, adaptive thinking, and new safety classifiers similar to Claude Fable 5.1. The model is suited for software development and knowledge work. Users need an active AWS account with Amazon Bedrock access and specific tools like Python 3.10+ and the Anthropic SDK to get started.

Sources

• We Need Google Trends for AI
• Voters are turning to AI on politics. Is it up to the task?
• Ex-DraftKings staff say AI targeted free bets to gamblers most likely to keep losing
• Saying hi to HAI: Exploring AI and the humanities at Stanford
• The Emerging M&A Map For AI Agent Security
• AI power is falling into the hands of ‘techno-oligarchs’, Spain's Sánchez warns
• Xi and Trump Seek Safe AI Without Slowing the Race for Supremacy
• Understanding Jensen Huang's AI optimism and how it influences President Trump
• Foreign Secretary Address to the UNSC on Artificial Intelligence
• Claude Opus 5.5 is now available on AWS