Thinking about machine empathy (1)
Amanda Ruggeri reports that many researchers question whether artificial intelligences can have empathy, even in principle. And if one were to produce something that looks like empathy, some think that wouldn’t be real. But isn’t this the same assumption about human uniqueness that can be applied to machine intelligence? The Turing test 鈥� in which a machine tries to pass as a person 鈥� wasn’t designed as an actual assessment, but as a thought experiment to undermine the assumption that a machine couldn’t be intelligent. Surely it also works in relation to empathy (9 March, p 32). After all, we don’t really know what is going on in an empathising brain. Couldn’t it be something rather like a large language model?
Thinking about machine empathy (2)
Many years ago, I wondered if there was any aspect of the human condition without a downside. I thought I had a winner in empathy. Then it became apparent that, for con artists, empathy is their most important weapon. This trait is the most powerful tool humans have. It therefore becomes our greatest exploitable weakness. Empathy-enabled AIs could be more dangerous than guns, knives or charismatic political and religious leaders. Handle with care!
Thinking about machine empathy (3)
We have no solid basis for the common assumption that other people are conscious, let alone that a therapist feels with us and isn’t merely emulating empathy.
Thinking about machine empathy (4)
Ruggeri clearly and correctly characterises empathy as “Ugh, I know exactly how you feel”. She goes on to say that “ultimately, you can’t really know what sadness is unless you have felt sad”. There is the rub: a truly empathetic therapist, friend or confidant needs to have really been there. Where? Where no AI has ever been.
Early concern over sports concussion was ignored (1)
Graham Lawton’s welcome piece on sports-related concussion recalled for me my earliest days in the UK’s National Health Service. In the early 1960s, I began working for a consultant physician who had a much-articulated hobbyhorse concerning brain injury, with particular reference to boxing and to football heading. He advocated as widely as he was able the need for changes in both sports, to outlaw heading the ball and to make blows to the head in boxing as unlawful as blows below the belt. Needless to say, his efforts of more than six decades ago went largely unheeded. He used to illustrate his point by saying: “If you fill a tin bucket with eggs, and then set about belting the bucket with a lump hammer, you would be a fool not to expect a few of the eggs to crack, at the very least (9 March, p 36).”
Early concern over sports concussion was ignored (2)
Anyone who has looked at the paths of seismic waves from an earthquake will have seen the waves refracted by the increasing densities of the crust, mantle and core. Multiple reflections back and forth are seen to occur, during which the waves repeatedly hit Earth’s surface. In various places inside Earth, these waves come to a focus where they cross over.
This raises the question of what happens in the brain when someone is struck on the head. Assuming a similar mechanism, even a small bump could cause some damage by displacing brain tissue where waves focus. This has obvious repercussions for any contact sport.
Still in harmony with Pythagorean music
In defence of Pythagoras, that he showed us music was based on harmonic intervals was a very impressive analysis with the primitive tools at his disposal. What is more, some of nature’s most wonderful singers, the Australian pied and grey butcherbirds, create beautiful melodies based on those intervals, and these can be transcribed using Western musical notation (9 March, p 15).
That isn’t to suggest that non-Pythagorean intervals can’t be enjoyable. There are examples in Western music: Scottish bagpipes aren’t tuned to the conventional scale. But Pythagoras laid the groundwork, and butcherbirds prove him right. They have been around longer than him.
Getting habituated to the dire state of the world
I liked your interview with Tali Sharot on habituation and found a lot of sense in it. In particular, it reminded me of how the endless stream of charity adverts on TV, many of which are repeated in graphic detail ad nauseam during the daytime to get repeat payments from senior citizens, can only have the effect of numbing us to the injustices they are trying to address (2 March, p 40).
Why can't water firms remove microplastics?
You report that boiling tap water can remove microplastics. But this only really works for hard water, as the microplastics co-precipitate with the calcium carbonate. In which case, you must also avoid using the water in the very bottom of the kettle. I think I would prefer a water company to supply water without microplastics (9 March, p 19).
One size won't fit all for long covid treatment
I feel a significant flaw in research on the use of exercise for post-viral conditions such as long covid is that many people aren’t well enough to participate. I have moderate ME and yet wouldn’t be able to take part in an exercise programme like the one in the article. Unfortunately, this can lead to a self-selecting sample of those who are moderately to mildly affected by such an illness. Findings can’t be generalised to those who are severely ill (17 February, p 14).
In some places, staking might be right for a tree
Just following up on James Wong’s look at the science of staking trees, one of the ever-present hazards for plants in urban environments is being knocked into, bashed and generally destroyed by passers-by, council mowers and the like (20 January, p 44). Might it be that tree staking succeeds in these environments purely by putting a physical barrier around the young plant?