GPT-4 vs. Llama 2: Which AI is Spreading More Misinformation Ahead of 2024 Elections?
Portfolio Pulse from Ananya Gairola
A study by Arthur AI has evaluated the misinformation spread by AI models from companies like Meta Platforms Inc., OpenAI, Anthropic, and Cohere AI. The study found that OpenAI's GPT-4 model hallucinated significantly less than its predecessor, GPT-3.5, and Meta's Llama 2 model. Cohere AI disputed the results, asserting the effectiveness of their retrieval automated generation technology for providing verifiable citations.
August 18, 2023 | 7:59 am
News sentiment analysis
Sort by:
Ascending
NEGATIVE IMPACT
Alphabet Inc.'s CEO Sundar Pichai admitted that AI platforms generally struggle with 'hallucination problems'.
The CEO's admission could potentially affect investor confidence in Alphabet's AI capabilities, leading to a negative impact on its stock.
CONFIDENCE 90
IMPORTANCE 60
RELEVANCE 60
NEGATIVE IMPACT
Alphabet Inc.'s CEO Sundar Pichai admitted that AI platforms generally struggle with 'hallucination problems'.
The CEO's admission could potentially affect investor confidence in Alphabet's AI capabilities, leading to a negative impact on its stock.
CONFIDENCE 90
IMPORTANCE 60
RELEVANCE 60
NEGATIVE IMPACT
Meta's Llama 2 AI model was found to have a higher tendency for misinformation compared to OpenAI's GPT-4.
The study's findings could potentially harm Meta's reputation in the AI industry, leading to a negative impact on its stock.
CONFIDENCE 90
IMPORTANCE 70
RELEVANCE 80