Connect with us

Hi, what are you looking for?

Tech & Science

New warning about the false data produced by large language models

These chatbots can be used to provide information, but it may be tainted by errors, or made-up or false information.

Microsoft plans to offer AI training to three million people in Japan
Microsoft plans to offer AI training to three million people in Japan - Copyright AFP Philip FONG
Microsoft plans to offer AI training to three million people in Japan - Copyright AFP Philip FONG

The world’s people increasingly rely on large language model (LLM) chatbots such as ChatGPT or Copilot to receive and organize information. There is growing concern over how these chatbots often make mistakes or provide  made-up or false information (which data scientists have taken to calling “hallucinations”).

An LLM algorithm functions by scanning enormous volumes of text to learn which words and sentences that frequently appear near one another and in what context. LLMs can be adapted to perform a wide  range of tasks across different domains and they produce forms of synthetic learning.

These technologies are said to hallucinate because they are built on problematic data sets or incorrect assumptions. Hence, the growing popularity of LLMs has raised concerns about their accuracy. These chatbots can be used to provide information, but it may be tainted by errors, or made-up or false information (hallucinations) caused by problematic data sets or incorrect propositions made by the model.

The questionable results produced by chatbots has led to growing disquiet among users, developers, and policy makers. The author argues that policy makers need to develop a systemic approach to address these concerns.

The current piecemeal approach does not reflect the complexity of LLMs or the magnitude of the data upon which they are based, therefore, the author recommends incentivizing greater transparency and accountability around data-set development.

A new research paper provides one of the first analyses of the many issues in data collection for LLM Chatbots and how governments are dealing with these e-challenges. In the paper, the author argues that these chatbots are creating disquiet because of how they are designed.  In AI, there is a significant focus on privacy or IPR related to chatbots by the news and press.

The paper is titled “Data Disquiet Concerns about the  Governance of Data for Generative AI” and it reviews the process and the specific problems in design. The paper has been produced by the Centre for International Governance Innovation.

Looking at the paper, Libza Mannan, Communications Advisor at the Centre for International Governance Innovation (CIGI), a Canada (Waterloo)-based think that works on issues at the intersection of technology and international governancey, tells Digital Journal that the paper is of significance to the business and academic community.

Mannan points out that the paper “Is one of the first analyses of the many issues in data collection for LLM Chatbots and how governments are dealing with these e-challenges.”

Mannan says: “The author argues that these chatbots are creating disquiet because of how they are designed.  In AI, there is a significant focus on privacy or IPR related to chatbots by the news and press. This paper reviews the process and the specific problems in design.”

The paper argues that policy makers need to develop a systemic approach to address these concerns.

Avatar photo
Written By

Dr. Tim Sandle is Digital Journal's Editor-at-Large for science news. Tim specializes in science, technology, environmental, business, and health journalism. He is additionally a practising microbiologist; and an author. He is also interested in history, politics and current affairs.

You may also like:

World

“What’s the point?” is a question Russia should have been asking for 120 years. It’s about time for an answer.

Business

A Togg electric car rolling off the assembly line in Gemlik near Bursa in western Turkey - Copyright AFP MOHD RASFANMathieu RABECHAULT and Anne...

World

The AfD bans its leading candidate Maximilian Krah from EU election campaign events over a series of scandals - Copyright ROSCOSMOS/AFP/File HandoutHui Min NEOGermany’s...

Tech & Science

housands of tech enthusiasts filed into Europe's self-declared biggest startup event VivaTech in Paris on Wednesday.