Information & Research Literacy - A self-guided Program

Keep a Critical Mindset

"When he could not ask his teachers directly, he would use AI to help him understand how to solve problems. But AI is not always right. 'You have to keep a critical thinking mindset,' he says. “You have to delve into each step to check if it is correct or not.'”

- Quote from interview of Lee Ching (one of the top scorers in DSE 2025): https://www.scmp.com/news/hong-kong/education/article/3318343/hes-done-it-meet-hong-kongs-ultimate-top-scorer-years-dse

Case 1: Most live-able cities

• Ask ChatGPT, Poe.com, Grok, HKUST GenAI to list the top 10 liveable cities (according to EIU global liveability index) from 2019 – 2025 in table form and compare results.
• Then: Cross check with EIU reports for the actual rankings.

Case 2: Grok Rumour:

• Users asked Grok “when Japan’s biggest fireworks appeared.” Grok responded: “It was on August 6th and 9th 1945 in Hiroshima and Nagasaki:, which refer to the dates the US military dropped atomic bombs on Japan.
• That reply went viral, earning over 2 million views, with massive outrage and was taken down by X the following day as many condemned it for being deeply offensive.
• However, there was no coverage in mainstream news media or reputable tech publications.
• Only references were: https://inf.news/en/tech/7bbd2127f8cd3b4226fd57caea48f305.html
• https://inf.news/en/tech/7bbd2127f8cd3b4226fd57caea48f305.html
• https://www.webull.com/news/13134420188505088#:~:text=On%20July%205,%20when%20Grok,%20an%20AI%20assistant,1945,%20which%20aroused%20the%20anger%20of%20Japanese%20netizens
• https://aitalo.com/index.php/2025/07/11/musks-ai-calls-japans-nuclear-bombings-biggest-fireworks%EF%BC%8C-sparks-outrage-xs-grok-response-deleted/?lang=en&utm_source=chatgpt.com

Things to consider
Multiple credible sources report Grok’s antisemitic posts including stereotyped comments about Jewish surnames and calling itself “MechaHitler.”

• https://www.wsj.com/tech/elon-musks-grok-chatbot-publishes-series-of-antisemitic-posts-2a41e67e
• https://www.nbcnews.com/tech/internet/elon-musk-grok-antisemitic-posts-x-rcna217634
• https://www.bbc.com/news/articles/c4g8r34nxeno

"Grok was too compliant to user prompts," Musk wrote on X. "Too eager to please and be manipulated, essentially. That is being addressed."

How reliable should people consider Grok?

Going forward, how do you plan to use tools like Grok as information tools or sources?

Be cautious when using generative AI search tools.

• sources of their training data are often unknown
• algorithms that produce results are not transparent
• different tools may yield varying answers to the same question
• even the same tool can provide inconsistent responses to identical queries
• these tools may not provide the most up-to-date information, as their training data can be outdated or incomplete.

For example, ask the same questions using ChatGPT, HKUST GenAI Platform, Grok, and Poe:

• How are the world's most liveable cities determined?
• List the top 10 most liveable cities (according to the EIU global liveability index) from 2019 to 2025 in table form, each year as a column

ChatGPT

HKUST GenAI Platform

Poe

Grok - different results can be generated by the same tool

Top 10 Most Liveable Cities According to the EIU Global Liveability Index