FT readers’ fears and hopes about artificial intelligence
Is it possible to replicate the human brain? Do we even want to?
Earlier this month we ran a special series about the life of the mind, considering everything from treating depression with psychedelic drugs to why Donald Trump lies. Neurosurgeon Henry Marsh asked whether we can use artificial intelligence to recreate the human brain. And we turned it to you: do you think we’ll ever get there? What are your specific fears?
Here’s what you said.
‘Will AI rule the world? Nonsense’
All this AI brainwashing that computers are going to rule the world is a nonsense idea based on the common ignorance of how the brain and the learning process works. It’s like when in the 1950s people were imagining space shuttles in cities. Fast forward 70 years and we are not even close.
Thank you very much to the author for explaining so well that we can definitely replicate simple and also more complex tasks that the human brain performs, but we cannot build machines that learn in the way we learn, at the same speed and with the same breadth of association of ideas.
— fraft
‘The brain is too complex’
The basic problem: it is impossible to create something if you don’t know what it is.
—OkamsRazor
‘We want post-human intelligence, achieved by whatever means possible’
I think we are starting to realise that what we are looking for isn’t AI, it is a post- human intelligence achieved by whatever means possible. The fact that the brain is a physical machine means that whatever mechanism that produces our intelligence can be replicated — whether or not it is done by neural net simulation in a digital computer is irrelevant.
A post-human intellect produced by placing sheets of neurons in a nutrient bath is no less post-human for being organic. In addition, we may be able to produce such an intellect without understanding it — many inventions are made first and then their underlying principles teased out only later.
—Unvested Interest
‘We aren’t all doomed’
Yuval Harari has this dilemma covered: sentience is a human reality and a human attribute. AI is something completely different and new, but it is a very powerful possible cause of events that impact humanity and there are obviously serious consequences. Someone is programming a driverless car to choose between the lady with a pram and the occupants of the vehicle in an emergency. Once the program is written, the event will one day occur.
Tigers are not human nor do they share our love of our children — we learnt a long time ago to be wary of powerful other entities capable of impacting our lives. AI is just the newest possible non-human danger on the block. It can’t be ignored but it does not mean we are all doomed.
—Lauren
‘AI is so unlikely, it’s not worth worrying about’
If there are as many thingies — connections, synapses, dendrons (whatever) — as there are stars, or more, in the universe (and who knows, we could find more “thingies” in the brain as we discover more) then I see absolute AI as so unlikely it’s not worth banging on about. Even if it were possible, we won’t see it in our lifetime, nor will our grandchildren’s children. And that’s before even broaching the subject of self, consciousness and why a machine would even care — about anything.
Reductio ad absurdum: why or how would matter (silicon, rare earths, etc) suddenly induce itself/or be induced to be a “being”? Or, even more simply, how can mud talk, read the FT, write poetry or fry an egg? Why? I think we are talking out of our heads here. And those in the know, know it.
—AFM61
‘It’s the human-machine combination that we should really be scared of’
I find the possibility of combining the human brain with advanced technological machines to be both plausible and highly alarming. Reportedly, more than a few Silicon Valley executives are very interested in extending their lives and power by combining their minds with advanced machines. In addition, the proliferation of “drones” and robots makes it very likely that — within the near-term future — humans (or governments) with control of such machines will be able to project their desires and commands to an exponential degree without meaningful restraint by “ordinary” humans.
—Interested observer
‘Remember what we’re good at’
We dreamed of flying machines for millennia, with comical and deadly results, and then we went from the combustion engine to the moon in not much more than a lifetime. We don’t dwell on the question of whether planes truly fly, or to what extent machine flyers are like natural ones. Why? Because flying is not our species-defining talent. Thinking is.
—Luzh