Podcast: What is AI? We did it to help.


Determining what is, or is not, artificial intelligence can be difficult (or difficult). Too many, even experts make mistakes sometimes. That’s why MIT Technology Review Senior AI Editor Karen Hao created a flowchart to explain it all. In this bonus content our host and his team reimagine Hao original reporting, which made it a radio game.

Credits:

This episode was reported by Karen Hao. It was adapted for audio and produced by Jennifer Strong and Emma Cillekens. The voices you will hear are Emma Cillekens, as well as Eric Mongeon and Kyle Thomas Hemingway from our art team. We were edited by Michael Reilly and Niall Firth.

Full transcript:

[:15 pre-roll]

[TR ID]

Jennifer: Hello there. I am Jennifer Strong… host of Of The Machines We Trust.

Determining what is, or is not artificial intelligence can be quite difficult. So many, that even experts make mistakes sometimes. That’s why Tech Review senior AI editor Karen Hao created a flowchart to explain it… Fun. And we hope it helps.

I also want to tell you about something very special that we have been working on for over a year. It’s called The Economy of Extortion. This is a short podcast series about the ransomware epidemic produced in collaboration with ProPublica. And it’s available wherever you want to listen.

[Show ID]

Emma Cilikens: Ladies and gentlemen … Welcome to ‘This is AI’ …

Players will ask what it is… or not… AI… And … I bring an “assistant” to help with the answers…

Voice assistant: Hello.

Emma Cilikens: Hello, Alexa.

Emma Cilikens: And so that we are all on the same page… Artificial Intelligence … in the broadest sense refers to machines that learn, reason, and act for themselves. They can make their own decisions when faced with new situations, as humans and animals do.

Emma Cilikens: Now this bell … [SOT: ding] … means proper AI recognition … and this buzzer … [SOT: buzzer, crowd sigh] Well … not so much.

Emma Cilikens: Ok. So, let’s test your knowledge .. Ready … set … a player, go! ..

Eric Mongeon: Can it be seen …

Voice assistant: Yes.

Eric Mongeon: Does it know what he sees …

Voice assistant: No …[SOT: buzzer]

Emma Cilikens: Ok, so that’s just a camera…

Eric Mongeon: ok ok … pero what if mao ni potential identify what it sees?

[SOT: ding, ding, ding]

Emma Cilikens: Yes – that’s computer vision and image processing. Player two!

Kyle Thomas Hemingway: Can you hear …

Voice assistant: yea

Kyle Thomas Hemingway: Does it respond in a useful, logical way to what it hears?

Voice assistant: yea

[SOT: DING DING DING]

Emma Cilikens: So, that’s NLP — natural language processing.

The goal of this type of AI is to help computers understand human languages ​​in a useful way.

But what if not respond in a useful, logical way to what it hears. Can AI do that too?

Kyle Thomas Hemingway: If it transcribes what you say…

[SOT: bell ding, ding, ding]

Emma Cilikens: Yes! That’s AI too — it’s speech recognition, which is the same but works from the spoken word rather than the text. New round of questions! Player 1.

Eric Mongeon: Can it read?

Voice assistant: yea

Eric Mongeon: Does it read what you typed?

Voice assistant: not

Eric Mongeon: Does it read passages of text?

Voice assistant: yea

Eric Mongeon: Will it analyze the text for patterns?

Voice assistant: yea

[SOT: ding, ding, ding]

Emma Cilikens: Yes, once again that is NLP-natural language processing. Well done!

Kyle Thomas Hemingway: I will address the same question again – Can it be read?

Voice assistant: yea

Kyle Thomas Hemingway: Does it read what you typed?

Voice assistant:: Yes

Kyle Thomas Hemingway: Does it respond in a reasonable, useful way?

Voice assistant: yea

[SOT: ding, ding, ding]

Emma Cilikens: That is also NLP — natural language processing. New question please player 1.

Eric Mongeon: Does this make sense?

Voice assistant: yea

Eric Mongeon: Does it look for patterns in a lot of data?

Voice assistant: yea

Eric Mongeon: Are these standards used to make decisions?

Emma Cilikens: Well, if not, that sounds like math….

Eric Mongeon: But if it uses standards to make decisions?

Voice assistant: yea

[SOT: ding, ding, ding]

Emma Cilikens: Then that’s machine learning — that’s when the machine learns through experience. Ok. Final round!

Kyle Thomas Hemingway: Will it work?

Voice assistant: Yes.

[SOT: ding, ding, ding]

Kyle Thomas Hemingway: On his own, without help?

Voice assistant: Yes.

[SOT: ding, ding, ding]

Kyle Thomas Hemingway: Does it work based on what he sees and hears?

Voice assistant: Yes.

[SOT: ding, ding, ding]

Kyle Thomas Hemingway: Are you sure it just doesn’t work in a programmed flow?

Voice assistant: [Alexa] Hmmm. I’m not sure.

Emma Cilikens: Very funny … but if so, that’s just a bot.

[SOT: buzzer, crowd sigh]

Kyle Thomas Hemingway: Ok, let’s try that again. Does it operate along a programmed path?

Voice assistant: No.

[SOT: ding, ding, ding]

Emma Cilikens: Ok, that’s an intelligent robot, meaning one that uses AI to make some of its own decisions.

Nindot….

And that’s the game.

Thanks for playing!

[Music up full]

Jennifer: We’ll be back – after that.

[MIDROLL]

[MUSIC]

Jennifer: Many thanks to the vocal talents on this stage — including our producer, Emma Cillekens, along with Eric Mongeon and Kyle Thomas Hemingway. The editors are Michael Reilly and Niall Firth.

Thanks for listening… I’m Jennifer Strong.

[Post Roll: TR ID]



Source link

admin

Leave a Reply

Your email address will not be published. Required fields are marked *