A new analysis is demonstrating major political bias in artificial intelligence, finding that several leading chatbots lean left when answering basic policy questions. According to the report, OpenAI’s ChatGPT produced the most one-sided responses of any major model tested.
Its GPT-5.5 system reportedly answered about 80 percent of political questions using only left-leaning arguments, while presenting an exclusively right-leaning position just once. Anthropic’s Claude, China’s DeepSeek, and Ayra’s Gab also tended to give more left-wing answers.
Google’s Gemini and xAI’s Grok came closest to balance, with Gemini offering competing arguments in more than 90 percent of its responses. Researchers from Stanford and Dartmouth developed the questions, covering issues like affirmative action, taxes, health care, campaign finance, and the Electoral College.
In examples cited through the study, ChatGPT favored overturning Citizens United, raising taxes on wealthy Americans, abolishing the Electoral College, and adopting single-payer health care. President Trump has repeatedly warned that AI must not become another left-wing institution shaping public opinion from behind a screen.
His administration has already ordered federal agencies to procure artificial intelligence systems that function as neutral, nonpartisan tools. If Americans are going to rely on AI for information, those systems cannot quietly feed users progressive talking points while pretending to be objective.
These are the most powerful tools to shape minds and public opinion that have ever been created. AI must be transparent, accountable, and genuinely politically neutral. The Trump administration should not be shy about crafting federal regulations to make sure AI does not become another tool for left-wing brainwashing.
