17 Comments

This is great, Bob. I guess it leads to fewer people “writing” ad copy, but if that is soul-sucking work anyway, maybe that is for the best. That monument ad is fantastic (don’t let time erase...). But I suppose there is still a craft to ad writing that AI can’t approach. I’m thinking of Don Draper in Mad Men, and that iconic scene, “The Carousel,” which you can see on Vimeo. Your opening example of reporters interviewing a candidate, however, made me wonder how many speech writers will be replaced by AI, and if AI might, in that way, gain some power over actual policy. It seems disturbing that AI weighs all opinions equally -- there appears to be no ability to curate knowledge, or to identify falsehood. There aren’t even “alternative facts” -- there is just a sea of words, and they’re all given equal credence. What we often want in leadership is discernment and moral authority. But that also requires discernment on the part of the people choosing the leader, and if we don’t know whether the words coming out of a candidate’s mouth are their own or bot-generated, I guess I can’t decide if that could be good for democracy (fueling a more aware and invested citizenship) or if, as I suspect, our default to laziness will make us perpetually more vulnerable to...what, exactly, I’m not sure. Another piece that came to mind while reading this is Ellen Ullman’s “Dining with Robots,” which takes some pains to show that a robot will never enjoy a fine dinner, a la Julia Child. The essay begins with a trip to a farmer’s market, that mecca for foodies, and ends with a trip to the supermarket. But Ullman reflects, in that robot-like grocery store, that instead of us making robots in our image, they have steadily been remaking us in theirs.

Expand full comment

Powerful, thank you!

Expand full comment

The most eye opening thing in Robert's piece is that AI did a very solid job of identifying what the public arguments are for private school vouchers AND how weak the public arguments are for supporting public education. AI will continue to amplify that difference until liberals start to do a better job of developing the social movement to support smart government, starting with public schools. And thus is the reason that Republicans have gone after public schools. They know where the power to block them lies and they know how to use the tools of the 21st century and how to recruit the leaders of the 21st century to advance their agenda.

Expand full comment

Interesting way to identify the public "knowledge" on a topic. I might have to try this to learn about water quality "improvements."

Expand full comment

I couldn't help but notice that both AI presidents were male and white. Imagine that - technology with bias built in!

Expand full comment

Yes. This will draw in the past rather than envisioning a future.

Expand full comment

Yes. I tried it a couple of more times and it came out the same...

Expand full comment

To the relief of writers everywhere, AI isn't very good (yet!) at answering open-ended questions. But drill down a little bit, and it's is really something. For example, see what might happen if you resubmit the request with a bunch of specifics like this:

"Give me 5 pros and 5 cons of Iowa's Education Savings Accounts program for a mother of 3 children under the age of 10 living in Exira. It needs to be less than 750 words and informational in nature. Start each pro with the word "pro," and each con with the word "con," and then provide a stat or data point to back it up."

I'm not sure how I feel about any of this yet. Using it as something like a research assistant seems ethically solid, as does your example of writing ad copy allowing the sales reps to focus on more important aspects of the job. But after that?? I dunno.

What I do know is that from a strictly technological standpoint, it's amazing. .

Expand full comment

Fascinating - but that music was very strange!

Expand full comment

Very, very strange...

Expand full comment

Clever, fun, fascinating. Also a bit terrifying (in fact, a macro-agg--where was my alert?) I'll be O.K., I'm sure, but may need to have Cherie screen your stuff for me for awhile.

Expand full comment

Surely you jest. Just wait until my next one!

Expand full comment

Interesting. Wonder what Ai would do if you asked it to write a pro and con about Ai. Keep up the writing by you.

Expand full comment

On a limb, here...I think we can agree that our political process is demented, or at least an augmented reality enhanced by some people who are a bit unbalanced. We need something to break through the cloudiness of parochialism. Hence, I see a future for this technology in giving credence to more reliable decision making, free from transitory, vacuous political fads when writing legislation. We are not there, yet...but when the wealth of human understanding and compassion become a defensible legislative action, maybe we can lean on AI-assisted legislation to draft 'for the greater good.'

Expand full comment

I hope you are correct, but this assumes AI can distinguish between good and evil. Interestingly, there are ethics built into it of a sort. For example, it won't let you write an anonymous love letter.

Expand full comment

Thanks for investigating. Big change for sure. Some good some very scary. Beth’s point on bias is right on.

Expand full comment

Wow! Great way to explore what this new technology might mean to all of us, in our own jobs and communities. More big change is clearly coming in media — gulp!

Expand full comment