ECONOMY & WORK
MONEY 101
NEWS
PERSONAL FINANCE
NET WORTH
About Us Contact Us Privacy Policy Terms of Use DMCA Opt-out of personalized ads
© Copyright 2023 Market Realist. Market Realist is a registered trademark. All Rights Reserved. People may receive compensation for some links to products and services on this website. Offers may be subject to change without notice.
MARKETREALIST.COM / NEWS

Here’s How AI Chatbots ‘Hallucinate’ And Why It Is Dangerous For Users

Turns out, AI chatbots are also capable of hallucinating and they do so by making up false information in response to a user’s prompt.
PUBLISHED DEC 29, 2023
Cover Image Source: Pexels | Photo by cottonbro studio
Cover Image Source: Pexels | Photo by cottonbro studio

The word “hallucination” is commonly associated with people hearing sounds and seeing things that no one else seems to hear or see. However, since the advancement of AI chatbots, the world has taken a different stand. Turns out, AI chatbots are also capable of hallucinating and they do so by making up false information in response to a user’s prompt. Further, “hallucinate” in the AI sense, is Dictionary.com’s word of the year as it best represents the potential impact that AI may have on “the future of language and life.”



 

An AI hallucination occurs when a large language model (LLM) generates false information. LLMs are AI models that power the commonly known chatbots like ChatGPT and Google Bard.

Photo illustration of the welcome screen for the OpenAI
Photo illustration of the welcome screen for the OpenAI "ChatGPT" app | Getty Images | Photo by Leon Neal

An AI chatbot hallucinates when users provide a prompt that is either made up or about something that the LLMs have no understanding of. In this case, the LLMs use statistics to generate false information in language that is grammatically correct and within the context of the prompt. Thus, the information appears plausible but it necessarily isn’t true. Sometimes AI hallucinations appear to be nonsensical but it is often hard to differentiate between true information and a hallucination.

One infamous example of an AI hallucination mentioned in a TechTarget report involved Google's chatbot, Bard. When the chatbot was prompted “What discoveries from the James Webb Space Telescope can I tell my 9-year-old about?" Bard responded with an untrue claim that the James Webb Space Telescope was the first to take pictures of an exoplanet outside this solar system. However, in reality, the first images of an exoplanet were taken in 2004 according to NASA, and it was not the James Webb Space Telescope as it was launched in 2021.



 

Despite that, Bard's answer sounded plausible and was consistent with the prompt. This poses a risk to users as someone who is not looking at the information with a skeptical eye may believe that the chatbot’s reply is true.

Photo illustration of the OpenAI ChatGPT's  answer to the question
Photo illustration of the OpenAI ChatGPT's answer to the question "What can AI offer to humanity?" | Getty Images | Photo by Leon Neal

There is no clear way to determine the exact causes of AI hallucinations on a case-by-case basis. Further, an immediate risk that AI hallucinations pose is that they significantly hurt user trust. With the development of advanced chatbots, users begin to experience AI as more real and develop more inherent trust. However, false information given out in hallucinations can corrode that trust leaving users with the feeling of betrayal.

Another problem with AI hallucinations is the lack of awareness of the problem. Users can be fooled with false information and this can even be used to spread misinformation, fabricate citations and references, and even be weaponized in cyberattacks. Another challenge is mitigating the AI hallucinations as it can be difficult or in many cases impossible to determine why the LLM generated the specific hallucination. Further, there are limited ways to fix these hallucinations as going into the model to change the training data can use a lot of energy.

Both OpenAI and Google have warned users about the mistakes that their AI chatbots can make. The firms have advised users to double-check their responses.

Image of Person's hand holding an iPhone using the Google Bard generative AI language model (chatbot) | Getty Images | Photo by Smith Collection
Image of Person's hand holding an iPhone using the Google Bard generative AI language model (chatbot) | Getty Images | Photo by Smith Collection

Tech organizations are also working on ways to reduce hallucination, as per a CNBC report. Google handles hallucinations through user feedback. When Bard generates an inaccurate response, users are encouraged to click the thumbs-down button and describe why the answer was wrong to help the chatbot improve. Further, OpenAI has implemented a strategy called “process supervision” in which the AI model rewards itself for using proper reasoning to arrive at an output instead of just rewarding the system for generating a correct response to a prompt.

MORE ON MARKET REALIST
The fans took issue with the clue being too tough and unnecessary for the contestants.
5 hours ago
Cuban wanted to strike a deal with the founder of Coconut Girl and gave her an ultimatum.
7 hours ago
Although the word did make a lot of difference, some fans felt that the decision was ridiculous.
8 hours ago
The guest, who found the item online, didn't think it was the real deal until he heard its value.
1 day ago
The player Samantha secured the first big win of Ryan Seacrest's run on the show.
1 day ago
The sisters wanted to keep their father's legacy alive, but also had their careers to focus on.
1 day ago
The shopper noticed that the watermelon which she bought three days ago was deflated.
1 day ago
The guest couldn't believe that her collection was worth so much money.
2 days ago
While all the Sharks were out, Mark Cuban saw the potential in "Garage Celebrations."
3 days ago
The loss was ironic as the retired teacher coulnd't get a phrase commonly used in classrooms.
3 days ago
The recalls were issued for a range of products including tires, air conditioners and power banks.
3 days ago
The guest who got the item as a gift from a neighbor had little to no idea about its significance.
3 days ago
Harrison got a good deal for the fossil although it wasn't what he had thought.
4 days ago
The player, Beth Barbee left everyone in the dust with her impressive puzzle solving skills.
4 days ago
Given her celebrity status, Harrison cut her some slack in the negotiations.
4 days ago
The deal ultimately fell apart as the seller, Scotty was looking to get $375,000.
5 days ago
The guest was blown away by the six-figure appraisal for his father's John Falter illustration.
5 days ago
Harvey just couldn't believe that the contestant could come up with an answer like that.
6 days ago
Antoinette's win was even more special because she needed a car more than anything at that point.
6 days ago
The player, Catrice Sandt, nearly blew the Bonus Round puzzle by saying too many words in the end.
7 days ago