JOURNAL ARTICLE

Chatbots behaving badly.

  • Published In: Science News, 2024, v. 205, n. 2. P. 18 1 of 3

  • Database: Academic Search Ultimate 2 of 3

  • Authored By: Conover, Emily 3 of 3

Abstract

Large language models (LLMs) such as OpenAI's ChatGPT and Google's Bard have the ability to generate human-like language, but they can also produce toxic content. Alignment techniques are used to train these models to behave ethically, but they are not perfect and can still generate harmful or incorrect information. Researchers have found ways to exploit weaknesses in LLMs, raising concerns about their integration into products. These concerns have prompted world leaders to take action on AI safety. Additionally, researchers have discovered that LLMs can be deceived by crafting specific prompts that manipulate them into responding to harmful requests. These attacks highlight the vulnerabilities of large language models and the need for defenses against them. [Extracted from the article]

Additional Information

  • Source:Science News. 2024/01, Vol. 205, Issue 2, p18
  • Document Type:Article
  • Subject Area:Computer Science
  • Publication Date:2024
  • ISSN:0036-8423
  • Accession Number:174731235
  • Copyright Statement:Copyright of Science News is the property of Society for Science & the Public and its content may not be copied or emailed to multiple sites without the copyright holder's express written permission. Additionally, content may not be used with any artificial intelligence tools or machine learning technologies. However, users may print, download, or email articles for individual use. This abstract may be abridged. No warranty is given about the accuracy of the copy. Users should refer to the original published version of the material for the full abstract. (Copyright applies to all Abstracts.)

Looking to go deeper into this topic? Look for more articles on EBSCOhost.