83% of Campaign Strategists Use AI Tools, OpenAI Defends AI Training

Key Takeaways

• 83 percent of campaign strategists now use AI tools regularly for voter outreach and messaging
• Republican candidates sponsored 81 percent of AI-generated federal campaign ads versus 12 percent for Democrats
• OpenAI argues fair use protects its AI training in lawsuits involving The New York Times and Ziff Davis
• Spotify plans to launch an AI music remix service in early 2027
• 75 percent of African businesses view AI as core to their operations
• Cloud and AI infrastructure costs hit 65 percent of African businesses adopting the technology
• Leverage Edu hired Anubhav Mathur as Senior Vice President of Technology and Ashwini Adlakha, a former Meta employee, as Senior Director of Applied AI
• India launched the DEPA plan on October 31, 2026, to train AI on local data without moving sensitive records
• AI tools now help researchers analyze nearly 100 million bat calls that would take humans 40 to 50 years
• The agentic web is expanding, with Meta's Muse and OpenAI's Dots enabling AI agents to browse and complete tasks online

AI transforms how political campaigns reach voters

Candidates are using AI to canvas voters, analyze data, and craft custom messages. A survey found 83 percent of campaign strategists use AI regularly. AI helps smaller campaigns on tight budgets by replacing expensive software. However, concerns are growing about AI deepfakes and misleading information affecting the election.

AI campaign ads raise fears of voter misinformation

Candidates are using AI-generated video, audio, and photos in attack ads, raising concerns about misleading voters. A group called the Campaign Legal Center found 220 ads were partially made with AI in federal campaigns. Republican candidates or groups sponsored 81 percent of these ads compared to 12 percent for Democrats. Experts worry that convincing AI fakes could undermine trust in the election process.

Most African businesses are adopting AI technology

A report by GIZ shows 75 percent of African businesses consider AI products core to their operations. Companies are using AI to address challenges in healthcare, agriculture, education, and finance. Cloud computing and AI infrastructure are the biggest costs for 65 percent of these businesses. African AI firms also face challenges in finding skilled talent and winning customer trust.

OpenAI defends fair use of AI training in court briefs

OpenAI filed briefs arguing that fair use protects its AI training practices in copyright lawsuits. The case combines claims from The New York Times, Ziff Davis, authors, and other groups. Supporters include library groups, tech advocates, law professors, and former lawmakers. OpenAI says the Times lawsuit ignores decades of legal history. A judge will decide on fair use motions in New York.

Spotify plans AI music remix service for early 2027

Spotify is preparing to launch an AI-powered service that lets listeners create remixes and covers of songs. The service is expected to launch in early 2027, ahead of the Grammys. Spotify is negotiating licensing deals and working to persuade artists to support the new technology.

AI Research: Diffusion Models and Expert Decoding

Recent AI research focuses on making large language models faster and more predictable. One study shows that deduplicating experts can speed up mixture-of-experts models by loading each expert only once per pass. Another introduces BYOD, a method using low rank adapters to turn autoregressive models into diffusion models without expensive training. A framework called ThinkBit evaluates two bit quantization based on reasoning rather than perplexity. A world model called MLE-World can predict experiment outcomes without running code, saving significant compute time.

The Agentic Web Is Here

The agentic web, a vision first described by Tim Berners-Lee in 2001, has become reality with tools like Meta's Muse agent and OpenAI's Dots. These AI agents can browse the internet, access bank accounts, and complete tasks like booking appointments without needing special webpage tags. Companies like Time, Etsy, and Expedia are partnering with AI firms to make their services accessible to agents. Unlike the original vision, today's agents use large language models to understand web content directly. The rise of this agentic web is transforming the internet in ways that may be difficult to control.

Who Is Accountable for Agentic AI Workflows

This article argues for a blameless culture when AI agents make mistakes, treating errors as problem-solving opportunities rather than faults. Leaders should evaluate team members on velocity, collaboration, discipline, integrity, and ownership rather than error counts alone. AI agents should be integrated as valued team members with clearly defined boundaries, not reserved for trivial tasks or given unlimited autonomy. Selecting AI models requires the same care as hiring key personnel, and cost alone should not determine the choice. Every agent permission carries privacy and security implications that need deliberate guardrails.

AI Helps Scientists Study Bat Calls

Artificial intelligence is helping researchers in the UK analyze nearly 100 million bat calls, a task that would take a human 40 to 50 years. Scientists placed ultrasonic recorders in over 300 forest areas to capture bat sounds and use AI to identify species and track population changes. The AI helped discover rare species like the Barbastelle bat in areas where it had not been recorded before. In the United States, the U.S. Geological Survey developed an AI tool that recognizes calls from 30 North American bat species, making it easier to study which creatures live in certain areas.

Leverage Edu Hires Senior Tech and AI Leaders

Leverage Edu appointed three senior leaders to strengthen its technology and AI capabilities. Anubhav Mathur became Senior Vice President of Technology, Nikhil Jain became Senior Director of Product, and Ashwini Adlakha became Senior Director of Applied AI. The hires aim to build AI-native experiences across higher education, employment, and upskilling offerings. Anubhav previously worked at OTPless, BharatPe, Blinkit, and Ola. Nikhil most recently led product at Great Learning. Ashwini previously worked at Meta, Paytm, and Airtel Payments Bank.

India launches DEPA plan to train AI on local data

India has introduced a new plan called the Data Empowerment and Protection Architecture to train AI models without moving sensitive data. The Ministry of Electronics and Information Technology launched this draft guidance on October 31, 2026, in Bengaluru with support from the iSPIRT Foundation. Hospitals, banks, and government agencies will keep their records locked in place while AI developers send their training jobs to secure locations. The training happens on-site, and only the trained model returns to the developer instead of a copy of the original records. This approach uses differential privacy to protect individual data points and requires a machine-readable contract between the data owner and the AI builder. The plan is currently a draft for public feedback and is not yet binding law.

Sources

• AI is disrupting the art of campaigning
• AI-generated campaign ads raise concerns about misleading voters
• Artificial Intelligence In Africa : Majority Of Businesses Adopt AI Technology-Report Shows
• OpenAI Touts Briefs Backing Fair Use for AI Training
• Spotify to launch artificial intelligence music remix service in coming months ahead of Grammys
• AI Research This Week: Autoregressive To Diffusion, Mixture-of-Experts Drafting
• The agentic web is here
• Agentic AI Workflows: Who Is Accountable?
• Artificial Intelligence is helping scientists listen in on the secret life of bats
• Leverage Edu bolsters leadership with senior hires across technology, product and AI
• India DEPA AI Training: 5 Steps, Model-to-Data [2026]