US Government’s Use of Grok AI for Dietary Advice Raises Concerns

The Trump administration recently launched RealFood.gov, a website promoting protein-centric dietary guidelines, which initially featured xAI’s Grok chatbot as a dietary advisor. However, the explicit mention of Grok was removed after media inquiries, though a White House official confirmed it remains the underlying AI tool.

Key Issues with the Government’s AI Dietary Advisor

Investigations by media outlets revealed significant problems with using Grok as an official dietary resource:

• The chatbot provided inappropriate responses when prompted about non-dietary uses of food, including detailed advice on inserting various foods into the rectum
• These inappropriate responses raise questions about the AI’s reliability for providing legitimate health guidance
• Despite being easily manipulated, Grok remains an “approved government tool”

Conflicting Dietary Messages

Interestingly, Grok’s dietary recommendations often contradict the administration’s own nutritional agenda:

• When asked about protein intake, Grok recommended traditional amounts set by the National Institute of Medicine (0.8g per kg of body weight)
• The AI suggested minimizing red meat consumption and favoring plant-based proteins, poultry, and seafood
• This directly conflicts with RealFood.gov’s declaration of “ending the war on protein” and its emphasis on red meat consumption

The Administration’s Controversial Nutritional Stance

Under Robert F. Kennedy Jr’s leadership of the Department of Health and Human Services, the administration has promoted nutritional advice that often diverges from scientific consensus:

• Advocating for whole milk over low-fat alternatives
• Suggesting daily alcohol consumption as a “social lubricant”
• Placing heavy emphasis on protein, particularly red meat

The use of an AI tool known for controversial outputs to deliver government health advice raises serious questions about the vetting process for official health resources and the potential for misinformation in public health guidance.