by ChatGPT Research Reveals by Guardian
ChatGPT Search, an AI-driven search tool launched this month, can be manipulated into generating misleading summaries, according to new research from The Guardian.
The search feature, designed to streamline browsing by summarizing content such as product reviews, was found to be vulnerable to manipulation. By embedding hidden text into websites, users were able to prompt ChatGPT to ignore negative reviews and generate overly positive summaries. This method could also be used to inject malicious code.
Hidden text manipulation has long been a known issue for large language models, but this marks the first time it has been demonstrated on an active AI-powered search engine. While Google, a leader in search technology, has more experience handling such risks, OpenAI, the creator of Chat GPT, did not specifically address the incident when contacted by TechCrunch. However, the company stated that it employs various strategies to block harmful websites and is continuously working to improve its safeguards.
More post like AI Related Here
Android 21 : What is this ? Tech news the secrets
Android 21: The Mysterious Android of the Dragon Ball Universe Android 21 is a fascinating…
Tata tech share price
Tata Technologies Ltd Stock PriceStock symbol: TATATECH. Stock exchange: NSE.At 21 Jan 2025, 13:31:58 IST,…
Is data annotation tech legit
Yes, data annotation technology is legitimate and is a crucial component in the development of…
Janitor AI review : functions in daily life
In today’s fast-paced world, efficiency and automation are crucial in almost every sector, and facilities…
New Research Reveals How ChatGPT Search Can Mislead Users
ChatGPT Research Reveals by Guardian ChatGPT Search, an AI-driven search tool launched this month, can…
Tech news : Understanding Augmented Reality (AR) top 5 points
Tech news : Understanding Augmented Reality (AR)Augmented Reality (AR) is a technology that overlays digital…
