FUTURE SHOCK
This is the script for Future Shock which aired on BBC Front Row on 11/04/23.
If you want to listen along click here.
image by Midjourney
I’m sat in a coffee shop reading a book called Future Shock. It was written by this guy Alvin Toffler, in 1970. It’s about what happens to people, and societies when they’re overwhelmed by change. Reading this in 2023 is kinda mind blowing. Not only are his predictions scarily accurate, but he writes in a way that seems to sum up the feelings of right now, which is crazy for a book that’s 50 years old. Listen.
ALVIN TOFFLER
Future shock is the dizzying disorientation brought on by the premature arrival of the future.
He goes on..
ALVIN TOFFLER
The striking signs of confusional breakdown we see around us - the spreading use of drugs, the recurrent outbreaks of vandalism and undirected violence, the politics of nihilism and nostalgia - can all be understood better, by recognising their relationship to Future shock.
Do you ever feel like there’s too many new things to keep track of? If we go back just 15 years - we didn’t have Uber or Bitcoin, or iPads or Insta. Everything seems to be speeding up. And the outcome, is disorientation at a societal level. The outcome is Future Shock.
PART 1: AI WORDS
image by Midjourney
I grew up in the 90’s - before the internet was a thing. And in my life, there’s been a few times I’ve discovered something mind-blowing. Google Earth was one. Napster was another. This year my mind was blown again, by Open AI’s GPT-3, the precursor to something I’m sure you’ve heard mentioned again and again. Chat GPT.
Close your eyes and imagine you’re in an old library. The librarian is not just any librarian, this happens to be the most read person in the world. Now, that librarian might be the wisest person you’ll ever meet, but with millions of books published annually - put next to ChatGPT, I’m not sure if they compete.
ChatGPT is a large language model: an LLM. That means it’s trained on words. And it’s one of many AI systems in development. This one, essentially, is that librarian on steroids. And just like the best librarians, it doesn’t mind having a little chat with you.
I’m in my studio looking at the website chat dot open AI. I type “How many books are you trained on?”
It tells me it’s trained on millions of books. Hundreds of millions of websites. And it can recall any of them, in seconds. I ask it to summarise the book Future Shock, and it quickly does:
Future Shock is a groundbreaking exploration of the social.. yeah! I know what you’re thinking - Amazon can do that too, right? But one thing Amazon can’t do is context switch: So let’s ask it to summarise the book in the form of a Haiku.
And within seconds, it replies:
Pace of change so fast
Future Shock sets in at last
New ways to outlast.
Maybe that use case seems trivial to you. I mean it kinda is, but, it can do so much more. And the fact that the core of its methodology is simply predicting the next word it should write, kinda belies how remarkable this system is.
When I told it I wanted to make more interesting music videos for YouTube based around technology, it suggested I generate AI images and project them onto the wall behind me whilst I play piano. Not a bad idea. In the last few months I’ve used it constantly.
NIECE 1 (O.S.)
Uncle Paul, who would win in a battle between Christmas tree and Pikachu?
BENBRICK (O.S.)
Do you think that’s a good question?
NIECE 2 (O.S.)
Yea! Well, a Christmas tree has baubles and...
Yeah! Over Christmas I got my nieces using it too. Let’s move on to the bit I’m most excited about, which is..
PART 2: AI SOUND
image by Midjourney
I’ve rediscovered playing piano, and most of that is thanks to my teacher, Noriko. She’s been pushing me to find my own interpretation of whatever it is I’m playing. Like this piece by Rachmaninoff.
I’m gonna to try and explain something here: I believe in the search for truth within music. The search to unravel complex emotions and nuance with.. sound. Listen to this.
This is Kūchūsen by Yoko Kanno. I find myself returning to it again and again. I feel pain at how perfect it is, and pain knowing I’ll most likely never come close to writing something like this. But then again, I’m not sure if anyone will. I’m certainly not sure if AI will.
A couple months back David Guetta shared a new piece of music he’d made that employs AI to mimic Eminem.
AI EMINEM
This is the future rave sound, I’m getting lost in an underground.
For me, this was a really novel use of AI technology, I’m not sure if it was artistic in a pure sense, but it was interesting. And fraught with legal issues, i’m sure.
This clip is from 2012. It’s some chords I came up with that existed as a voice note on my phone for years. Last year we recorded at Abbey Road for Have You Heard George’s Podcast? and those chords were transformed by the BBC Concert Orchestra.
It took ten years from an initial idea to a realised version. And I think I needed that time. I’m not sure whether artificial intelligence solving everything immediately is necessarily good for the human condition.
Still, there are countless experimentations happening with sound. Like this one. This was made by a guy in Canada using two different AI’s - HarmonAI and Jukebox. The lyrics, music, and voice are all AI and I love it.
LYRICS
On the road the horns are blaring,
in the wind and rain, the lights are flaring,
It’s never going to be the same.
It was trained on 1.2 million songs, that were cut up into small clips. And from all that data it’s made a song. Imagine it re-recorded with a live jazz band. Imagine where we’ll be in five years.
PART 3: THE FUTURE
image by Midjourney
I need to say more about some of the issues surrounding AI. It’s able to write a paragraph or produce a song because it’s been taught. It’s been trained. The source of the data it’s trained on is really important.
And whilst an artist can use someone else’s work as inspiration, they can’t copy it. It’s the same for songs, and even if the AI is using the original work purely as quote inspiration, I still think artists should have a say if their content is being used to train these large AI models.
I recently asked ChatGPT for a quote on Ghana’s journey to independence for an episode I’m making with George The Poet. It provided the perfect Eisenhower quote, but when we tried to locate the source, or even better an audio clip, we couldn’t find one. Luckily, with ChatGPT you can ask it to explain what it does, and so I asked it if the quote was verbatim.
And within seconds, it replies: no.
This is truly worrying. I’ve been thinking about it for days now. A lie can be halfway around the world before the truth has it’s shoes on.
One of my favourite podcasts is called All In. Four mega successful tech entrepreneurs and thinkers. In this episode they’re talking about AI and their takes are really interesting. This is David Sacks.
DAVID SACKS (O.S.)
This is the power to rewrite history, it’s the power to rewrite society, to reprogram what people learn and what they think. This is a god like power.
And that’s what really worries me. Cos what if we didn’t have time to research that Eisenhower clip I mentioned earlier, so we made a version using a model of his voice? Who would fact check it? Or is it all fake news going forward? As Orwell said “The past was erased, the erasure was forgotten, the lie became truth.” When I think about AI I’m excited, but I can’t lie: there’s a sense of dread there.
Later in that episode of All In, Sacks says, “The government is not smart enough to regulate AI yet” And this is certainly true, whichever country you live in.
Let’s go back to the book I mentioned at the start of this piece. In its intro Toffler says “No serious futurist deals in predictions.” That’s funny to me, because he goes on to make a series of predictions again and again.
So in a bid to show that I’m also not a serious futurist: I’m going to leave you with a predictions of my own: I believe within the decade we’ll see a government run by AI. And instead of voting for a particular political party, you’ll vote for a particular algorithm.
See you in 2033 then?
