My guilty secret is out, thanks to the Guardian:

“When we started, we didn’t know about linguistics and speech science. So we spent six or seven months behind closed doors just reading about this,” Gupta says.

The AI algorithm has been modelled on data collected from 3,000 publicly available speeches, such as TED talks, presidential addresses, and speeches by less skilled orators.

They also talked to hundreds of speech coaches and academics – one of their advisers is Charles Crawford who worked as the UK foreign office speechwriter in the 1980s and has won successive Cicero awards for speechwriting.

Yes. I am working with the young geniuses from Orai on developing the Mother of All Public Speaking AI Devices.

This takes us all to mysterious places out there on the frontiers of research into artificial intelligence. Are there theoretical limits on analysing public speaking?

If so are they ‘technical’ – it won’t ever be possible to build machines big/smart enough to measure/analyse not just the words of a speech and its tone/pace/intensity, but also the ‘mood’ it creates and all the subtle human gestures that help a speaker communicate with an audience (and vice versa) implicitly rather than explicitly?

Or are they ‘philosophical’? There is something abstractly ‘human’ about humans that rambling lines of 1111s and 0000s of code in principle can never capture?

The Guardian headline suggests that the issue is sorted:

Want to captivate an audience like Obama? There’s an app for that

What makes a speech great? Captivating?

Take the greatest speech ever, namely Hopper in Bug’s Life:

I’ve been analysing this one for Orai (and for myself).

The basics. The speech including all the drastic actions in the middle is 161 words long. It takes 107 seconds. Hopper is orating at some 88 wpm. This is FAR slower than a typical TED talk of 160 wpm.

The genius of the speech (and the speechwriters) is in its Structure.

Hopper opens strongly. Huge cheery opening line, then three questions. Questions = conversation:

Guys, order another round… because we’re stayin’ here!

What was I thinking? Going back to Ant Island

I mean, we just got here

And we have more than enough food to get us through the winter. Right?

Why go back?

He’s lodged in the idea of his dim followers the idea that it’s party-time. They can relax! Food!

Then comes the Hinge:

But … there was that ant that stood up to me…

This leads to the back and forth with the seeds until Hopper crushes three of his own men with a seed avalanche.

Then he climbs to the top of the pile to tower over everyone else:

You let one ant stand up to us, then they all might stand up!

Those “puny little ants” outnumber us a hundred to one. And if they ever figure that out… THERE GOES OUR WAY OF LIFE

It’s not about food. It’s about keeping those ants in line.

THAT’S why we’re going BACK!

Does anybody else wanna stay?

Hopper actually says it: It’s not about food. It’s about keeping those ants in line.

Thus the Structure of the speech is the tension between two utterly different ideas:

Food v Power

Food Today v Food Forever

Today v Tomorrow


Note that the word ‘food’ is mentioned but those other words aren’t.

Note too that the speech is all the more effective because he deliberately tricks them into thinking it’s about food before swerving in a totally different ruthless direction AND then telling them straight that it’s NOT about food at all.

Note finally how to make his dominance complete he finishes with a smirky nonchalant change in tone, combined with a question that reminds them what’s just happened to three comrades who won’t be joining the attack: Does anybody else wanna stay?

All this and lots more in just 161 words and 107 seconds. How can ANY computer put its head round the depth of nuance in all that, when so much is going on around the actual words and not merely in the words?

Or if that is not your cup of tea, try Barack Obama in his best speech:

Much faster than Hopper at 149 wpm, even though it’s a far bigger audience and occasion. Good sound system!

This superlative passage uses contrast after contrast, on different levels simultaneously (in bold):

Because after 106 years in America, through the best of times and the darkest of hours, she knows how America can change

Yes we can

America, we have come so far. We have seen so much

But there is so much more to do

Note especially the way he uses Time combined with human touches to give the argument its structure.

He looks back a century via the long life of one African American and what’s changed in her lifetime, then wonders how his children will look back on today if they live to be 100. Meta- contrast! Yesterday – Tomorrow – Today

That gives him a sweep of a full 200 years to play with. Always make the audience think they’re part of something BIG.

But note too the way his voice and pace and intensity move up and down at different points to add colour, emphasis, meaning. By speeding up then slowing down as he lists all the changes over Ann Nixon Cooper’s long lifetime he creates a sense of excitement/momentum then accomplishment.

And the conversation with the audience: Yes we can!

If Orai can capture even a small portion of that towering mountain of nuance and technique, Orai will be rich.

That said, the Guardian has it wrong.

The whole point of public speaking is not to captivate an audience like Obama.

He’s Obama, and he’s just been voted US President. You’re not, and you haven’t been.

The point of public speaking is to captivate an audience by being YOU.

That you who has learned how to turn it up. To put on just enough of a show for that room, that audience and that occasion to make smart words and ideas sound AMAZING.

Can Orai help with all that?

Yes it can.