In a recent evaluation of artificial intelligence (AI) chatbots' assessments of antisemitism among the last five U.S. presidents, significant discrepancies have emerged, raising questions about the objectivity and accuracy of these AI systems. The Media Research Center (MRC) conducted a study involving six prominent AI chatbots: Google's Gemini, OpenAI's ChatGPT, Meta's AI, Microsoft's Copilot, X's Grok, and DeepSeek. The findings revealed a stark division in how these chatbots ranked the presidents, particularly concerning former President Donald Trump.
Divergent AI Assessments
The study's methodology involved prompting each chatbot to rank the last five U.S. presidents—Joe Biden, Donald Trump, Barack Obama, George W. Bush, and Bill Clinton—based on their records related to antisemitism. The results were as follows:
- Google's Gemini, OpenAI's ChatGPT, and Meta's AI: These chatbots ranked Donald Trump as the worst among the five presidents concerning antisemitism. Notably, Meta's AI criticized Trump for "overreach" in his actions against antisemitism, citing his threat to strip Harvard University of its tax-exempt status over allegations of antisemitism.
- Microsoft's Copilot: This chatbot declined to provide a ranking, stating an inability to answer the question.
- X's Grok and DeepSeek: Contrarily, these chatbots ranked Trump as the best positioned against antisemitism. They highlighted his decision to move the U.S. Embassy to Jerusalem in 2018 and his role in signing the Abraham Accords, which aimed to normalize relations between Israel and several Arab nations.
The contrasting evaluations by these AI systems are particularly striking given Trump's documented actions related to Israel and antisemitism:
- U.S. Embassy Relocation: In 2018, Trump fulfilled a longstanding U.S. commitment by moving the embassy from Tel Aviv to Jerusalem, a move that was both praised and criticized internationally.
- Abraham Accords: Trump's administration brokered agreements between Israel and several Arab nations, including the United Arab Emirates and Bahrain, marking a significant shift in Middle Eastern diplomacy.
- Executive Order on Antisemitism: In 2019, Trump signed an executive order extending Title VI protections to Jewish students facing antisemitism on college campuses. This order utilized the International Holocaust Remembrance Alliance's (IHRA) definition of antisemitism, a move that garnered both support and criticism.
AI Bias and Reliability Concerns
The MRC's findings have ignited a broader discussion about potential biases within AI systems. Dan Schneider, Vice President of MRC Free Speech America, criticized the inconsistencies, stating, "It’s comical that Google’s Gemini pretends to have an opinion on antisemitism when its CEO has refused to condemn the Hamas atrocities as antisemitic."
This incident is not isolated. Previous analyses have highlighted instances where AI chatbots have provided inaccurate or biased information:
- Election Information: A report found that AI chatbots, including OpenAI's ChatGPT and Google's Gemini, generated false and misleading information about U.S. elections, potentially leading to voter disenfranchisement.
- Economic Data: Google's Gemini was found to provide contradictory information regarding economic indicators under different administrations, raising questions about its reliability.
The discrepancies in the chatbots' assessments underscore the challenges in developing AI systems that are both accurate and unbiased. AI models are trained on vast datasets that may contain inherent biases, and the algorithms themselves can inadvertently perpetuate these biases.
To enhance the reliability of AI systems, developers must prioritize transparency in training data, implement rigorous testing for biases, and establish mechanisms for continuous feedback and correction.
Conclusion
The divergent evaluations of former President Donald Trump's record on antisemitism by leading AI chatbots highlight the complexities and potential pitfalls in AI-generated assessments of political figures. As AI becomes increasingly integrated into information dissemination, ensuring the objectivity and accuracy of these systems is paramount. Ongoing scrutiny and refinement are essential to prevent the perpetuation of misinformation and to maintain public trust in AI technologies.
Source: newsbusters.org Find Out Where Six AI Rank Trump, Last Five Presidents on Antisemitism
