We asked ChatGPT to write a layoff memo and it turns out this is 'a difficult time'

• The recent layoff memos Google, Amazon, and Meta put out have tended to follow predictable patterns.
• Insider asked ChatGPT, the AI chatbot, to draft a layoff memo for a pretend tech company, Gomezon.

Read enough layoff memos and you'll start to notice some patterns.

There's the high-level nod to challenging macroeconomic conditions: high inflation, the pandemic, and economic uncertainty — as though the rest of us live under a rock; the cursory information about severance packages; and sometimes, but not always, a mea culpa.

There's nothing more predictable, however, than the memos' semi-tortured tone about how "difficult" it is to make cuts.

"I have some difficult news to share." That's how Sundar Pichai, the CEO of Google's parent company, Alphabet, began his memo this month announcing that 12,000 employees were losing their jobs. Andy Jassy, the CEO of Amazon, said he'd made the "difficult decision to" cut more than 18,000 employees in his early-January memo about a round of layoffs. Meanwhile, the CEO of Meta, Mark Zuckerberg's, November memo described cutting 11,000 employees as "some of the most difficult changes" his team has made in the company's history.

Granted, there's no easy way to deliver this news. And stripping your employees of their livelihoods is, no doubt, difficult. But please CEOs, spare us the hand-wringing. Workers deserve more than a clichéd condolence card.

It makes you wonder: Did all these leaders go to the same CEO school where they learned to write generic messages dismissing large numbers of employees? Or did they all commission the same robot to write their memos for them?

Perhaps.

Insider asked ChatGPT, the viral AI chatbot sweeping the internet and the new go-to tool for just about every writing task, to whip up a layoff memo for a pretend tech company, Gomezon. It took a bit of prodding, but ChatGPT completed the assignment and acknowledged that laying off employees is indeed "difficult."

It's 'inappropriate' for AI to handle sensitive matters, per ChatGPT

Open AIs' ChatGPT has its fair share of problems. The chatbot is rife with bias, spreads misinformation, allows students to cheat, and can also enable questionable business practices.

But it's also pretty handy. It can dash off a passable resignation letter and cover letters good enough to fool recruiters. It can whip up news articles, offer solid investment advice, and even write software code.

Surely it could compose a layoff memo of reasonable quality, right?

Not so fast, apparently. ChatGPT labeled my first request "inappropriate."

That wasn't the answer I'd expected, but in some ways it made me feel somewhat better about the technology. For all the bellyaching about the dark side of AI, which is legitimate, ChatGPT's refusal to write the memo suggests to me that the chatbot "gets it" on some level.

ChatGPT is right: If layoffs must happen, companies must do them with the "utmost care and consideration," and communicate with "compassion and empathy."

A sample template that reeks of false sincerity

With a second request, however, ChatGPT relented. The chatbot stipulated that its memo was only a template and needed to be tailored to fit my make-believe company.

It's not surprising that ChatGPT's sample memo was alarmingly similar to the ones we've read recently. Which is to say that the memo was stiff and reeked of false sincerity.

Maybe I should have accepted the technology's initial reluctance to write the layoff memo. It's a job for a human, not a chatbot.

ChatGPT, apparently, has principles. I was the one who forced the issue, making for a "difficult time."