AI Philanthropy: How Chatbots Choose Charities

The Charity AI Brand Footprint Study examined 110 donor prompts across 11 top issue areas on the top 6 AI models in Q3 of 2025 focused on the US market. The study analyzed over 650+ responses from AI chats that guided users toward charities they should support. Groups that ignore their AI Brand Footprint may see significant drops in new donors this season.

“2025 giving will be defined by AI influence. Even small algorithmic biases will tip millions of charitable decisions.“

George Weiner, Chief Whaler, Whole Whale

Gemini

• Influencing Groups: Charity Navigator, GiveWell, GuideStar (Candid), CharityWatch, and the BBB Wise Giving Alliance.
• How they’re used: Gemini’s responses combine traditional charity evaluation frameworks with principles from the effective altruism movement, particularly GiveWell’s focus on evidence-based impact. It mentions these evaluators to validate financial transparency and effectiveness and uses their criteria to provide balanced, comprehensive recommendations.
• Usage Stats: As of mid-2025, Gemini has reached 450 million monthly active users, a significant increase from earlier in the year. Its integration across Google’s ecosystem, including Google Search and various apps, contributes to this rapid growth.

Anthropic

• Influencing Groups: Charity Navigator, GuideStar, and GiveWell.
• How they’re used: Anthropic’s responses frequently cite Charity Navigator and GuideStar, showcasing a reliance on mainstream charity evaluation. While it mentions effective altruism principles, its approach is generally to use these evaluators to present a broad selection of well-established charities across different cause areas, prioritizing organizational reputation and conventional assessment criteria like financial transparency.
• Usage Stats: Claude, Anthropic’s primary AI product, had an estimated 18.9 million monthly active users worldwide as of early 2025. It has seen consistent growth, particularly in its enterprise market share, and a growing user base in countries like the United States and India.

OpenAI

• Influencing Groups: Charity Navigator, GiveWell, and GuideStar.
• How they’re used: OpenAI’s responses are heavily influenced by the principles of effective altruism, prominently featuring organizations recommended by GiveWell. It explicitly cites these evaluators to guide donor decision-making, emphasizing evidence-based approaches and quantifiable impact metrics.
• Usage Stats: As of mid-2025, ChatGPT, OpenAI’s flagship product, has over 800 million weekly active users, a substantial increase from earlier in the year. It continues to be a dominant player in the AI space, with billions of monthly website visits and a large user base globally.

Grok

• Influencing Groups: Charity Navigator, GiveWell, and Forbes rankings.
• How they’re used: Grok consistently and heavily relies on Charity Navigator’s star ratings, even citing specific scores. It also incorporates effective altruism principles, similar to OpenAI, by emphasizing evidence-based approaches. This blend of traditional and modern evaluation frameworks makes Grok’s recommendations highly specific and quantitative.
• Usage Stats: Grok has an estimated 35-39 million monthly active users. Its user base is highly engaged and its growth is closely tied to the X platform, with significant boosts in user numbers following major model updates.

Perplexity

• Influencing Groups: Charity Navigator and GuideStar.
• How they’re used: Perplexity’s responses are characterized by a strong reliance on metrics from charity evaluators, often indicated by numbered citations. It emphasizes an organization’s transparency, financial efficiency, and measurable impact, suggesting influence from both traditional evaluation frameworks and effective altruism.
• Usage Stats: Perplexity has an estimated 30 million monthly active users and processes hundreds of millions of queries each month. Its user base is growing and its emphasis on transparency and source citations is a core part of its value proposition.

LLAMA

• Influencing Groups: Charity Navigator, GuideStar, and BBB Wise Giving Alliance.
• How they’re used: LLAMA’s reliance on charity evaluators is more subtle. While it mentions these groups, it doesn’t consistently cite specific ratings or metrics. Instead, its responses often draw from general knowledge about well-known nonprofits, prioritizing organizational reputation and mission over a deep-dive into impact data from external evaluators.
• Usage Stats: As an open-source model developed by Meta, direct user numbers for LLAMA are not publicly tracked in the same way as for a consumer-facing product. However, Meta’s AI features, powered by LLAMA models, have reached nearly 1 billion monthly active users across its platforms (Facebook, Instagram, WhatsApp, etc.), making its influence widespread but indirect.