Urgent.News

What's breaking now, across thousands of outlets.

AI

SUSS drops AI detector as more Singapore universities question reliability of such tools

NUS, SUTD and SIT do not use automated AI detectors, while SMU says such tools are only one input when investigating possible misconduct.

SUSS drops AI detector as more Singapore universities question reliability of such tools

The Singapore University of Social Sciences (SUSS) has disabled its university-wide AI writing detector from August 11, acknowledging the limitations of such tools in determining whether generative AI was used inappropriately. This decision does not affect SUSS's expectations for students to uphold academic integrity and use generative AI responsibly.

The university emphasizes that AI writing detectors can estimate the likelihood of AI-generated text but cannot conclusively determine misconduct. NTU and SMU have also announced plans to discontinue or reduce the use of AI detection tools, recognizing their inherent unreliability. While SMU continues to use AI detection tools, it stresses that they are only one input among several in investigating potential misconduct.

NTU and SUSS maintain that not using generative AI is not considered academic misconduct and that students are responsible for the accuracy and integrity of their submissions.

Written by urgent.news from CNA - Singapore's reporting — not their text. Machine-written — may contain errors; check the original before relying on it.

Also reported by 1 other outlet

Read the original at channelnewsasia.com →

More in AI

jina-embeddings-v4 as an OpenAI-Compatible Embeddings Server

jina-embeddings-v4 is a self-hosted server for the jina-embeddings-v4 embedding model with an OpenAI-compatible /v1/embeddings endpoint. It runs on a single NVIDIA GPU.

  • jina-embeddings-v4 is open-source, self-hosted server for embedding model
  • Compatible with OpenAI's embeddings service request/response format
  • Deploy on single NVIDIA GPU with 10GB+ VRAM, no per-request charges

New advancements in Generative AI

New advancements in Generative AI If you spent last year fine-tuning LLMs just to get a JSON payload back without markdown backticks wrapping the entire thing, you probably noticed the goalposts…

  • Shift from chatbot API to agentic workflows
  • Structured outputs with JSON schema support
  • Local execution of Llama 3.1 8B and Mistral 7B

More from Monday 24 August →