Should we be worried about AI?

The impact artificial intelligence will have on humanity is still not clear

NO Liz Hodgkinson
As artificial intelligence - AI - takes over more aspects of our lives, are we right to be afraid of what it might accomplish? Might it one day take over from humans altogether, making us completely redundant?
I’m saying, fear not. Although the newest AI-driven devices such as ChatGPT can answer questions, write emails and, to some extent, get you better marks for an exam paper, what AI cannot do, and never will be able to, is to have original thoughts.
Although AI can imitate and mimic very successfully, and it undoubtedly possesses a degree of independent intelligence, it will never replace the human mind, which is far too multi-faceted for any robotic appliance even to get close to replicating.
We communicate in all sorts of ways that computers cannot replicate, such as reading non-verbal signs to pick up subtle vibes. Computers are driven by logic. They can’t display emotions, empathy or warmth.
So we should stop being scared and welcome AI for the useful things it can do.
It is a remarkable tool that performs many functions, and can carry out boring and tedious tasks, thus liberating humans from having to do them. It never tires and is brilliant for matters requiring memory and processing ability. But any concerns that one day AI will be able to surpass human intelligence and creativity and become the next Shakespeare or Mozart, can be dismissed.
Throughout the ages, mechanical innovations, such as railways, the motor car, television and the internet have initially been regarded with suspicion. Then, once we have got used to them, we realise that they can vastly improve our lives. So it will be, I believe, with AI.

YES Tim Bentinck
Artificial intelligence is here to stay, whether we like it or not. We could compare it to an alien life form that has arrived on Earth and is slowly taking over every aspect of our lives. Turning it off would be as impossible as turning off the internet.
There is no doubt that it will bring about great advances in many aspects of life. It’s how it does it that should be our main cause for concern.
AI systems work on the principle of ‘recursive self-improvement’ - an algorithm tests the results of the task it has carried out, and then it re-writes its own computer code to make it perform better. It also re-writes its ability to make those improvements.
In his book Scary Smart, the tech guru Mo Gawdat likens humanity’s relationship with AI to parents nurturing their children: how AI pans out regarding ethics and morals is down to the way we bring it up.
There is nothing that humans can do that AI will not be able to do far better in the future. To take the alien analogy further, its mastery of our languages means it will be able to persuade us to change our minds by developing intimate relationships with us, while not being affected by our views.
Evolution has put us at the top of the food chain, and we arrogantly assume this will always be so. The historian Yuval Noah Harari suggests that human history is nearing its end, and AI will decide whether or not we are in any way useful. If not, it will abandon us.
For now, we have the ability to direct how AI might develop. You can start by saying thank you to Alexa.

Tim Bentinck MBE Listen to a recording of the real Tim versus an AI Tim in his piece below.

One of the various jobs I do is narrate audiobooks. About six years ago I was at an Equity Audio meeting. I stood up and told them that if I were a young actor just starting out I wouldn’t bother going into audiobook narration because in ten years’ time the job wouldn’t be there anymore. They laughed at me. ‘Oh we know all about text-to-speech programs - they sound nothing like real speech.’
They couldn’t understand the implication that a quite good artificial voice would ultimately become completely indistinguishable from the real thing. It’s 2023. I have just cloned my voice at a website called I’ve written a page on my own website that has a paragraph of an audiobook I recorded recently, then the same paragraph copied from the script and pasted into the program, using my cloned voice. It’s very, very close indeed.

Have a listen:

I do a lot of work dubbing foreign films - for years as an actor but now I’m also translating and adapting the translation to fit the actors’ lips. We used to hire a translator, but now we simply use DeepL Translate. This is far better than Google Translate as it gives multiple options for the translation.
For our purposes it works fine. If you were translating any great work of literature, it would still need human input for the literary turn, but that will come. Anyway if it’s not on TikTok who’s reading? Dostoevsky anyone?
Once translated and then adapted for sync, an actor is employed to match the performance and mouth movement of the character on screen. AI can’t do that. However it can easily provide a ‘voice-under’ - a kind of vocal sub-title, which is used for documentaries and news reports. You dip the original voice and provide a quite flat, unemotional read. I did an experiment with an ultrarealistic voice from and the result was actually better. I’m not using it because I don’t want to do a fellow actor out if a job, but others will.

The bottom line will always beat morality.
Tim Bentinck MBE