×
Microsoft seeks patent to fix AI hallucinations
Written by
Published on
Join our daily newsletter for breaking news, product launches and deals, research breakdowns, and other industry-leading AI coverage
Join Now

AI hallucination mitigation: Microsoft’s patent proposal: Microsoft has filed a patent application for a method aimed at reducing or eliminating AI-generated falsehoods, addressing one of the most significant challenges in generative AI technology.

  • The patent application, titled “Interacting with a Language Model using External Knowledge and Feedback,” was submitted to the US Patent & Trademark Office (USPTO) in 2023 and made public on October 31, 2024.
  • The proposed method introduces a “response-augmenting system” (RAS) that enables AI models to automatically gather additional information based on user queries and assess the usefulness of their responses.
  • This system would allow AI chatbots to inform users when they believe their answer is incomplete or questionable, and users could provide feedback on the responses.

Key features of the proposed system: The RAS is designed to enhance the accuracy and reliability of AI-generated responses by leveraging external sources and user feedback.

  • The system can access information from online sources or datasets to verify and supplement its responses to user queries.
  • If the AI model’s answer doesn’t include relevant information available from external sources, the RAS can deem the response “not useful.”
  • The method doesn’t require developers or companies to fine-tune their existing models, potentially making it easier to implement across various AI platforms.

Current status and implications: While the patent application is still under review, it represents a significant step towards addressing the persistent issue of AI hallucinations.

  • The USPTO website doesn’t show a patent number for the application, indicating that it may still be in the review process.
  • Microsoft has not yet commented on whether this patent is related to their previously announced Azure AI Content Safety tool, which offers AI-powered verification for business chatbots.
  • The Azure AI Content Safety tool performs fact-checking on AI responses before delivering them to users, adding an extra layer of verification to ensure responses are “grounded” in existing data sources.

The broader context of AI hallucinations: Hallucinations, or false responses, remain a critical challenge for generative AI, affecting the technology’s reliability and usefulness across various applications.

  • Notable AI systems, including Google’s AI Overviews and X’s Grok AI, have produced significant errors, ranging from giving absurd advice to spreading misinformation about elections.
  • Even industry leaders like Apple’s CEO Tim Cook have acknowledged that their AI systems are not immune to hallucinations.
  • OpenAI’s Whisper audio transcription tool has recently been found to produce frequent hallucinations, raising concerns about its use in sensitive environments such as US hospitals.

Industry trends and energy considerations: Despite the challenges posed by AI hallucinations, major tech companies continue to invest heavily in AI infrastructure.

  • Tech giants like Google, Microsoft, and Meta are exploring nuclear power as a potential solution to meet the increasing energy demands of AI data centers.
  • This trend highlights the industry’s commitment to advancing AI technology while addressing associated challenges, including energy consumption and environmental impact.

Potential impact and future outlook: Microsoft’s patent application represents a promising step towards improving the reliability of AI-generated content, but challenges remain.

  • If successful, this technology could significantly enhance the trustworthiness and usefulness of AI chatbots and other generative AI applications.
  • However, the effectiveness of the proposed system in real-world scenarios remains to be seen, and its implementation across different AI platforms may present additional challenges.
  • As AI continues to evolve and integrate into various aspects of daily life and business operations, addressing issues like hallucinations will be crucial for widespread adoption and trust in AI technologies.
Microsoft Wants to Patent a Fix for AI Hallucinations

Recent News

AI-powered computers are adding more time to workers’ tasks, but there’s a catch

Early AI PC adopters report spending more time on tasks than traditional computer users, signaling growing pains in the technology's implementation.

The global bootcamp that teaches intensive AI safety programming classes

Global bootcamp program trains next wave of AI safety professionals through intensive 10-day courses funded by Open Philanthropy.

‘Anti-scale’ and how to save journalism in an automated world

Struggling news organizations seek to balance AI adoption with growing public distrust, as the industry pivots toward community-focused journalism over content volume.