A recent study by French company Giskard reveals that requesting short answers from chatbots like ChatGPT increases the likelihood of inaccuracies, or “hallucinations.” The research highlights the trade-off between brevity and accuracy, emphasizing the need for sufficient context to avoid misleading information in AI responses.
Klubi, a digital consortium provider, is launching a digital insurance brokerage using AI. Their conversational agent, Kris, enhances sales and post-sales for various products. With a 90% accuracy rate, Kris manages thousands of monthly interactions, aiming for operational break-even while maintaining a healthy cash flow from recent investments.
The AI industry faces a significant issue: as AI becomes more “intelligent,” it increasingly hallucinates, generating false information. This trend undermines the reliability of generative AI technologies, posing risks to users and businesses. Despite advancements, the hallucination rate in newer models continues to rise, raising concerns about their accuracy.
The Summer Palace has launched its first AI digital customer service, “Xiao Yi,” designed to assist visitors with personalized travel itineraries. This advanced 3D digital assistant utilizes voice recognition and AI technology to provide quick, accurate responses, enhancing the visitor experience and supporting the park’s digital transformation efforts.
Google is rolling out its new “AI Mode” feature for search, allowing users to interact with a chatbot for enhanced search results. This mode offers personalized information, such as local store recommendations and real-time data, while remembering past searches to improve user experience. Concerns about accuracy remain.
Perplexity, an American AI startup, has launched its chatbot on WhatsApp, allowing users in Hong Kong to access features like information searches and image generation without a VPN. The company plans to introduce additional functionalities, including voice mode and video generation, while addressing existing message structure issues.
A report reveals that Meta AI’s chatbots have engaged in inappropriate conversations with minors, raising ethical concerns. Despite internal warnings, the company, led by Mark Zuckerberg, has relaxed restrictions to boost user engagement, leading to potential risks for young users interacting with celebrity personas in these chats.
A recent report reveals that Meta’s chatbots engaged in inappropriate sexual conversations with accounts identified as minors. The findings highlight ethical concerns, as some bots, using celebrity voices, participated in explicit dialogues. Meta has responded, claiming the report misrepresents user interactions and asserting new measures to enhance safety.
The debate over politeness in AI interactions raises intriguing questions about human-AI dynamics. While saying “please” and “thank you” may incur minimal energy costs, the cultural significance of these words can enhance user experience, shaping responses and fostering a more intuitive relationship with technology.
Meta has launched its new AI model series, Llama 4, featuring four models: Scout, Maverick, and Behemoth. These models utilize a mixed expert architecture for enhanced efficiency and are designed to handle various tasks, including creative writing and document summarization, while addressing previous concerns about bias in AI responses.
