LANG 1407 - Academic English for Engineering Studies

Designed on Canva: https://www.canva.com/design/DAFLh4z-eQQ/view

• Published information is not necessarily trustworthy or useful.
• Always evaluate the credibility & suitability of information sources.
• In addition to RADAR, there are also other evaluation frameworks:

Cross-check and Fact Check

• Can you find the information, theories etc. backed up by other information sources (newspapers, magazines, scholarly journals)
• Cross-check statistics, news reports, etc.
• Beware of fake news circulating on social media, explore these fact-check sites
  • AFP Fact Check by The Agence France-Presse
  • AP Fact Check by The Associated Press
  • FactCheck.org® A Project of The Annenberg Public Policy Center
  • Google Fact Check Explorer
  • HKBU Fact Check by The Hong Kong Baptist University School of Communication

A common problem with many generative AI text tools is that they generate plausible but factually incorrect information (hallucinations) or fabricated references (hallucitations), examples:

• Hallucinations in ChatGPT: A Cautionary Tale for Biomedical Researchers
• Lawyer cites fake cases generated by ChatGPT in legal brief: The high-profile incident in a federal case highlights the need for lawyers to verify the legal insights generated by AI-powered tools

So, if you are using texts generated by ChatGPT or other generative AI tools, be cautious and cross check the information. Evaluate the information provided, just like anything you plan to use academically.

Just as we've covered in the sections above:

1. Cross-check what it says & do "lateral reading"

• Check against your own knowledge
• Check against Wikipedia, Britannica, and/or other reliable sources to see if they state the same information
• This is not new with generative AI, it's always been good scholarly practice

2. Confirm that any references it provides are real

• Check them against PowerSearch, Google Scholar, other Library databases, etc.

3. If the references (citations) are real, check that they support the claim

• This is *NOT* a new issue with AI tools, it's always been important to do so.
• This is proper scholarly practice when working with sources both in print and online.

Lateral Reading (3 min 33 sec) from the University of Louisville Library