I asked ChatGPT to write an obituary for Hazel McCallion. It did not go well.

Official portrait - Hazel McCallion

I wrote my first obituary for Hazel McCallion at least five years ago when the former Mississauga mayor was just 96.

It is a common practice for media outlets to assign reporters to research and pre-write obituaries for prominent public figures while they are still alive.

Indeed, detailed obituaries appeared online less than an hour after the official word of her passing on Jan. 29.

Premier Doug Ford announced McCallion’s death

The initial articles about her death were comprehensive because of this newsroom practice of having rough drafts on standby to be polished with details whenever someone’s death is confirmed (I’m not sure how much of my work made it into the coverage by my former employer).

The coverage was updated throughout the day as more officials and residents weighed in with their recollections of her life. The 101-year-old transformed the Toronto suburb from a sleepy bedroom community into Canada’s 7th largest city.

This made me wonder if ChatGPT could be used to help reporters create that initial rough draft. As an experiment, I typed “Write a 500-word obituary for Hazel McCallion.”

Screen capture of ChatGPT output

I thought I might catch it plundering language from Wikipedia or any of the many biographies detailing her 36-year-long career as mayor.

What it produced was worse that I could imagine. It was so riddled with errors it should serve as a warning for any journalist (or journalism student) hoping to substitute this for old-fashion research and note-keeping.

“Hazel McCallion, one of Canada’s most iconic and beloved politicians, passed away on January 30, 2023 at the age of 96. McCallion was the mayor of Mississauga, Ontario for an unprecedented 36 years, making her the longest-serving mayor in Canadian history. “

ChatGPT output

It failed the basics, getting both the day of her death wrong (January 29) and her age (she was 101). None of this was a great mystery to any local reporter who knew her or covered her. Indeed, her 100th birthday was a local news event.

What’s more, she is not Canada’s longest-serving mayor. That title goes to Milton’s Gordan Krantz. So much for OpenAI scraping Wikipedia as a source.

It only got worse from there.

McCallion was born in 1921 in Prince Edward Island and grew up in Montreal. She received her degree in commerce from McGill University and later went on to become a teacher. In 1968, she moved to Mississauga and soon became involved in local politics. She was elected to the Mississauga City Council in 1970, and just two years later, she was elected mayor.

ChatGPT output

None of this is accurate. She was not born in PEI. She did not attend McGill. She worked as an office manager for an engineering firm, not as a teacher. She and her husband moved to the Mississauga area in the 1950s.

The dates of her political career also bear no relationship to reality. Her political life started as a deputy reeve and later mayor of the town of Streetsville, which was incorporated into Mississauga in 1974. She served one term as a councillor before being elected mayor in 1978.

While she had many accomplishments as mayor, I don’t recognize any of the ones ChatGPT raised:

One of McCallion’s greatest accomplishments was her role in securing a major regional airport for Mississauga. She worked tirelessly to bring the airport to the city, and her efforts paid off in 1991 when Pearson International Airport opened its doors.

ChatGPT output

Pearson airport has been operating in one form or another since 1937. She had plenty of connections to the long-running airport ChatGPT could have highlighted, such as her fight to ensure it paid property taxes or her role on its board of directors.

I tried to give ChatGPT another chance by re-generating the results. Once again, it got her age and place of birth wrong, but then came up with this gem of an accomplishment:

She was a fierce opponent of the proposed development of a large airport on the city’s waterfront, which would have threatened the local ecosystem and wildlife.

ChatGPT output

As someone who’s grown up in the Greater Toronto Area, I have no idea what the AI writer is talking about. The only value to come from this exercise is perhaps in generating a document I can use to help student practice their fact-checking skills.

If anything, it points to a saving grace for journalists and writers who penned the reflections pieces about McCallion’s passing.

ChatGPT is no replacement for local knowledge and experience.

56 responses to “I asked ChatGPT to write an obituary for Hazel McCallion. It did not go well.”

  1. […] and were subsequently optimized for. This has resulted in an AI with a silver tongue but a “just sorta skimmed the Wiki before chiming in” aptitude of fact […]

  2. […] and were subsequently optimized for. This has resulted in an AI with a silver tongue but a “just sorta skimmed the Wiki before chiming in” aptitude of fact […]

  3. […] and were subsequently optimized for. This has resulted in an AI with a silver tongue but a “just sorta skimmed the Wiki before chiming in” aptitude of fact […]

  4. […] and were subsequently optimized for. This has resulted in an AI with a silver tongue but a “just sorta skimmed the Wiki before chiming in” aptitude of fact […]

  5. […] have been subsequently optimized for. This has resulted in an AI with a silver tongue however a “simply sorta skimmed the Wiki earlier than chiming in” aptitude of truth […]

  6. […] and were subsequently optimized for. This has resulted in an AI with a silver tongue but a “just sorta skimmed the Wiki before chiming in” aptitude of fact […]

  7. […] and were subsequently optimized for. This has resulted in an AI with a silver tongue but a “just sorta skimmed the Wiki before chiming in” aptitude of fact […]

  8. […] and were subsequently optimized for. This has resulted in an AI with a silver tongue but a “just sorta skimmed the Wiki before chiming in” aptitude of fact […]

  9. […] and were subsequently optimized for. This has resulted in an AI with a silver tongue but a “just sorta skimmed the Wiki before chiming in” aptitude of fact […]

  10. […] and were subsequently optimized for. This has resulted in an AI with a silver tongue but a “just sorta skimmed the Wiki before chiming in Advertisement ” aptitude of fact […]

  11. […] had been subsequently optimized for. This has resulted in an AI with a silver tongue however a “simply sorta skimmed the Wiki earlier than chiming in” aptitude of reality […]

  12. […] and were subsequently optimized for. This has resulted in an AI with a silver tongue but a “just sorta skimmed the Wiki before chiming in” aptitude of fact […]

  13. […] and were subsequently optimized for. This has resulted in an AI with a silver tongue but a “just sorta skimmed the Wiki before chiming in” aptitude of fact […]

  14. […] have been subsequently optimized for. This has resulted in an AI with a silver tongue however a “simply sorta skimmed the Wiki earlier than chiming in” aptitude of reality […]

  15. […] and were subsequently optimized for. This has resulted in an AI with a silver tongue but a “just sorta skimmed the Wiki before chiming in” aptitude of fact […]

  16. […] and were subsequently optimized for. This has resulted in an AI with a silver tongue but a “just sorta skimmed the Wiki before chiming in” aptitude of fact […]

  17. […] and were subsequently optimized for. This has resulted in an AI with a silver tongue but a “just sorta skimmed the Wiki before chiming in” aptitude of fact […]

  18. […] and were subsequently optimized for. This has resulted in an AI with a silver tongue but a “just sorta skimmed the Wiki before chiming in” aptitude of fact […]

  19. […] and were subsequently optimized for. This has resulted in an AI with a silver tongue but a “just sorta skimmed the Wiki before chiming in” aptitude of fact […]

  20. […] and were subsequently optimized for. This has resulted in an AI with a silver tongue but a “just sorta skimmed the Wiki before chiming in” aptitude of fact […]

  21. […] and were subsequently optimized for. This has resulted in an AI with a silver tongue but a “just sorta skimmed the Wiki before chiming in” aptitude of fact […]

  22. […] had been subsequently optimized for. This has resulted in an AI with a silver tongue however a “just sorta skimmed the Wiki before chiming in” aptitude of reality […]

  23. […] have been subsequently optimized for. This has resulted in an AI with a silver tongue however a “just sorta skimmed the Wiki before chiming in” aptitude of truth […]

  24. […] and were subsequently optimized for. This has resulted in an AI with a silver tongue but a “just sorta skimmed the Wiki before chiming in” aptitude of fact […]

  25. […] and were subsequently optimized for. This has resulted in an AI with a silver tongue but a “just sorta skimmed the Wiki before chiming in” aptitude of fact […]

  26. […] and were subsequently optimized for. This has resulted in an AI with a silver tongue but a “just sorta skimmed the Wiki before chiming in” aptitude of fact […]

  27. […] and were subsequently optimized for. This has resulted in an AI with a silver tongue but a “just sorta skimmed the Wiki before chiming in” aptitude of fact […]

  28. […] had been due to this fact optimized for. This has ended in an AI with a silver tongue however a “simply sorta skimmed the Wiki sooner than chiming in” flair of truth […]

  29. […] and were subsequently optimized for. This has resulted in an AI with a silver tongue but a “just sorta skimmed the Wiki before chiming in” aptitude of fact […]

  30. […] and were subsequently optimized for. This has resulted in an AI with a silver tongue but a “just sorta skimmed the Wiki before chiming in” aptitude of fact […]

  31. […] have been subsequently optimized for. This has resulted in an AI with a silver tongue however a “just sorta skimmed the Wiki before chiming in” aptitude of truth […]

  32. […] and were subsequently optimized for. This has resulted in an AI with a silver tongue but a “just sorta skimmed the Wiki before chiming in” aptitude of fact […]

  33. […] and were subsequently optimized for. This has resulted in an AI with a silver tongue but a “just sorta skimmed the Wiki before chiming in” aptitude of fact […]

  34. […] and were subsequently optimized for. This has resulted in an AI with a silver tongue but a “just sorta skimmed the Wiki before chiming in” aptitude of fact […]

  35. […] and were subsequently optimized for. This has resulted in an AI with a silver tongue but a “just sorta skimmed the Wiki before chiming in” aptitude of fact […]

  36. […] and were subsequently optimized for. This has resulted in an AI with a silver tongue but a “just sorta skimmed the Wiki before chiming in” aptitude of fact […]

  37. […] and were subsequently optimized for. This has resulted in an AI with a silver tongue but a “just sorta skimmed the Wiki before chiming in” aptitude of fact […]

  38. […] and were subsequently optimized for. This has resulted in an AI with a silver tongue but a “just sorta skimmed the Wiki before chiming in” aptitude of fact […]

  39. […] and were subsequently optimized for. This has resulted in an AI with a silver tongue but a “just sorta skimmed the Wiki before chiming in” aptitude of fact […]

  40. […] and were subsequently optimized for. This has resulted in an AI with a silver tongue but a “just sorta skimmed the Wiki before chiming in” aptitude of fact […]

  41. […] and were subsequently optimized for. This has resulted in an AI with a silver tongue but a “just sorta skimmed the Wiki before chiming in” aptitude of fact […]

  42. […] and were subsequently optimized for. This has resulted in an AI with a silver tongue but a “just sorta skimmed the Wiki before chiming in” aptitude of fact […]

  43. […] and were subsequently optimized for. This has resulted in an AI with a silver tongue but a “just sorta skimmed the Wiki before chiming in” aptitude of fact […]

  44. […] have been subsequently optimized for. This has resulted in an AI with a silver tongue however a “just sorta skimmed the Wiki before chiming in” aptitude of reality […]

  45. […] and were subsequently optimized for. This has resulted in an AI with a silver tongue but a “just sorta skimmed the Wiki before chiming in” aptitude of fact […]

  46. […] and were subsequently optimized for. This has resulted in an AI with a silver tongue but a “just sorta skimmed the Wiki before chiming in” aptitude of fact […]

  47. […] and were subsequently optimized for. This has resulted in an AI with a silver tongue but a “just sorta skimmed the Wiki before chiming in” aptitude of fact […]

  48. […] and were subsequently optimized for. This has resulted in an AI with a silver tongue but a “just sorta skimmed the Wiki before chiming in” aptitude of fact […]

  49. […] and were subsequently optimized for. This has resulted in an AI with a silver tongue but a “just sorta skimmed the Wiki before chiming in” aptitude of fact […]

  50. […] and were subsequently optimized for. This has resulted in an AI with a silver tongue but a “just sorta skimmed the Wiki before chiming in” aptitude of fact […]

  51. […] and were subsequently optimized for. This has resulted in an AI with a silver tongue but a “just sorta skimmed the Wiki before chiming in” aptitude of fact […]

  52. […] and were subsequently optimized for. This has resulted in an AI with a silver tongue but a “just sorta skimmed the Wiki before chiming in” aptitude of fact […]

  53. […] and were subsequently optimized for. This has resulted in an AI with a silver tongue but a “just sorta skimmed the Wiki before chiming in” aptitude of fact […]

  54. […] and were subsequently optimized for. This has resulted in an AI with a silver tongue but a “just sorta skimmed the Wiki before chiming in” aptitude of fact […]

  55. […] and were subsequently optimized for. This has resulted in an AI with a silver tongue but a “just sorta skimmed the Wiki before chiming in” aptitude of fact […]

  56. […] and were subsequently optimized for. This has resulted in an AI with a silver tongue but a “just sorta skimmed the Wiki before chiming in” aptitude of fact […]

Leave a reply to ChatGPT is suddenly everywhere. Are we ready? – TechCodex Cancel reply