US Congress Members Demand Answers from Google over ‘AI Overview’

Points

• US lawmakers, led by Representative Adam Schiff, demand information from Google CEO Sundar Pichai about the ‘AI Overview’ feature.
• Concerns raised over AI Overview outputting incorrect and potentially dangerous information.
• Lawmakers request answers to eleven questions about accuracy, safety, and misinformation mitigation.

A group of United States lawmakers, led by Representative Adam Schiff, has demanded information from Alphabet/Google CEO Sundar Pichai regarding the company’s “AI Overview” feature for Search. The lawmakers have set an August 1 deadline for Pichai to respond.

The AI Overview feature provides a summary window at the top of Google Search pages when a user makes a query. This summary usually includes a short text answer, an image if relevant, and supporting links. However, there have been numerous reports of AI Overview outputting incorrect and potentially dangerous information.

Cointelegraph

Lawmakers’ Concerns

The complaint, published by Schiff’s office and co-signed by four other members of Congress, outlines several issues with AI Overview. One major concern is the AI’s inability to distinguish between satire and factual information. For instance, AI Overview reportedly advised users to “eat at least one rock a day for vitamins and minerals,” citing the satirical news site “The Onion.” Other instances include the AI spreading debunked conspiracy theories, such as claiming Barack Obama was the first Muslim US president.

Google’s Efforts and Lawmakers’ Demands

While acknowledging Google’s efforts to combat misinformation, the lawmakers assert that more needs to be done to ensure the reliability and accuracy of information provided by AI Overview. The complaint requests answers to eleven questions related to the company’s efforts to improve accuracy and safety, fact-checking procedures, sourcing, transparency, and collaboration with authorities.

The Challenge of AI Reasoning

The core issue with AI systems like AI Overview is their lack of true understanding or reasoning. Despite advancements, AI still struggles with context and the ability to apply learned experiences to new situations. These limitations have been observed in earlier large language models, such as OpenAI’s GPT-2 and Meta’s defunct “Galactica” system, which was shut down due to backlash over its inaccuracies.

Conclusion

The demand from US lawmakers for more information about Google’s AI Overview highlights the ongoing challenges and responsibilities of AI in delivering accurate and reliable information. As AI technology continues to evolve, ensuring its safe and effective use remains paramount.

解説

• Accuracy and Safety: Ensuring that AI systems provide accurate and safe information is crucial for maintaining public trust. The issues raised by lawmakers highlight the need for continuous improvement in AI technology.
• Fact-Checking and Transparency: Implementing robust fact-checking procedures and ensuring transparency in AI-generated content are essential steps in mitigating misinformation.
• AI Limitations: The inherent limitations of AI in understanding context and reasoning underscore the importance of ongoing research and development to enhance AI capabilities.