Editor's pick: The meaning of life, artificial and observant
Teal Burrell discusses human purposes and goals – which we clearly have from biological evolution (28 January, p 30). That raises the question: what purpose would an artificial intelligence have? If its software does not have a built-in set of goals, would an AI develop one at random from its observations of the world around it, or would it just pull the plug on itself?
Editor's pick: The meaning of life, artificial and observant
Burrell mentioning that those with a strong sense of purpose had a better chance of survival in concentration camps reminded me of an article on Arctic winter stays in the 16th and 17th centuries (3 April 1993, p 38). Those with a strong record of religious observance fared much better. The authors say: “Celebrations were something to look forward to; they gave a feeling of unity and the brief illusion of being at home… Perhaps this emphasis on religious observance was less a sign of innate piety than an intuitive psychological insight on the part of the more successful leaders.”
An urgent need to grant inhuman rights, to what?
Recent commentary has raised the possibility that sapient AI, if developed, might be assigned the right to vote (17/24/31 December 2016, p 18) and might need to be freed from slavery as a form of suffering (Letters, 28 January). As we move towards the threshold where research might create an entity capable of suffering in a human-like way, we ought to establish safeguards to protect the rights of sapient AIs as research subjects from the very start.
Review mechanisms to safeguard such human rights are well established, ensuring adherence to the Declaration of Helsinki through research ethics boards and committees. Perhaps we need to consider what should trigger the involvement of these committees in advanced artificial intelligence research.
An urgent need to grant inhuman rights, to what?
Robert Willis suggests artificial intelligences need rights (Letters, 28 January). I turn on a light bulb. Does it feel warm? I turn it off. Have I hurt it? Did I hurt it by turning it on? I turn my 1980s programmable calculator on and off, and ask the same questions.
In the future, I turn on my brain-embedded, internet-connected, super-smart AI chip that the law mandated all children must have to eliminate education inequality. Have I created a “me” with two votes? If it disagrees with me and I turn it off, have I killed it or committed election fraud by preventing it from voting?
Will the law mandate that it is always on and so visible to the security services to keep us safe? If only this were science fiction.
First class post
Not poisoning bees would be so much better. Just saying
Alix about the worth of work on robot pollinators (18 February, p 14)
We can share data only when rules are respected
Your leader article criticised the PACE trial of treatments for chronic fatigue syndrome, which I co-led, for not sharing data (11 February, p 3). On behalf of my colleagues, I need to point out that this part of your commentary is misleading. The PACE team has shared data many times with researchers who have agreed to respect its confidentiality, including with a Cochrane Collaboration group who have undertaken a meta-analysis of exercise therapy (which is in review) using individual patient data. We have not voluntarily released data to the general public as we do not have our participants' consent to do this. We suggest that the consent of trial participants must be considered in any discourse about sharing trial data.
The editor writes:
• We do not criticise PACE, nor those controlling access to statins data. We do suggest that consent for future analyses should be considered during data collection, in the knowledge that such research is likely to need this.
Perhaps reality is a process in mind
I was fascinated by your recent article on the essence of reality (4 February, p 29). I am trying to understand the origins of behavioural difficulties in young children, and I see a connection.
Alfred North Whitehead, who co-wrote Principia Mathematica with fellow philosopher Bertrand Russell, eschewed any material reality. In 1929, he wrote: “There persists… the fixed scientific cosmology which presupposes the ultimate fact of an irreducible brute matter, or material… senseless, valueless, purposeless… It is this assumption that I call ‘scientific materialism’… which I shall challenge as being entirely unsuitable to the scientific situation.” He held that reality is a dynamic process.
In the 17th century, philosopher Gottfried Leibniz similarly believed that reality is “activity”.
Current systems theorists do not assume a bedrock material of “reality”. Measurement is an intrusion and represents a temporary state – in which not everything can be measured.
I find “” fascinating. For me it has required a complete suspension of traditional ways of thinking, and offered relief from the idea that behaviour has discrete, static , which leads to the blaming of children and families.
Is it time for a major rethink of the nature of “reality”?
Perhaps reality is a process in mind
Anil Ananthaswamy does not distinguish what is made (or thought) by humans from what is really out there. A few pages later, Frank Swain writes that it is impossible to “still believe we can take an objective view of nature, or that we are fundamentally separate from it” (4 February, p 42). Spot on. In the non-human universe there is no machinery, no writing – and there are no mental tools like dimensions, and no time. There is energy and motion, and because energy exists in different forms, there is space. Both are “there” irrespective of humans' conception of them.
Ananthaswamy mentions time as something we might think of as reality, along with particles, energy and space. Time is not a part of “reality”; it is a generalised measuring device thought up, albeit inadvertently, to measure motion. Equally, the idea that “information” is a part of reality, never mind its true bedrock, needs to be challenged.
Information, however it is defined, is a truly human concept and only part of the universe in that it is in the human mind.
Very many mansions in the multiverse
Shannon Hall quotes Michael Hall as saying that the main problem with Hugh Everett's many-worlds interpretation of quantum mechanics comes from confusion over what constitutes a measurement (21 January, p 28). In the well-known thought experiment, the decay of a radioactive atom determines a cat's fate. In the many-worlds interpretation, when the atom decays the cat dies, and a parallel universe is created in which it lives. But it seems to me that at every moment in which the cat does not die a parallel universe must be created in which it does.
If we assume time is quantised at the “Planck time” (5.391 × 10-44 seconds), then for each second the cat remains alive in our universe, 2 × 1043 new universes are created. The quantum multiverse must be very crowded.
An infinitely puzzling speed limit question
Michael Brooks tells us that “the ratio of the speeds of light and gravity rapidly went to infinity” after the big bang (26 November 2016, p 8). My dimly remembered high-school maths suggests that one of them must have gone to zero. That really would be strange.
The editor writes:
• It would have been more precise to say that the ratio rapidly approached infinity.
Welcome to the Valles Marineris Wildlife Park!
Sarah Moles notes that it is hard to decide which species to save with limited resources (Letters, 4 February). But if our longer term future includes terraforming other planets, then we will have to take other species with us as DNA, rather than as live and potentially dangerous fellow passengers. So the question for us now is: what to save that allows us to resurrect something in the future?
The options include saving a species in the wild, in a zoo, in frozen form or as a DNA sequence. The more of them we save, the more likely it is we could give each a place in future. But if we tried to resurrect an extinct mammal, would we be able to restore its genetic diversity, as well as its gut and skin microflora? And what about the rest of its ecosystem?
A way of seeing may help a diagnosis
As a scientifically trained adult with autism, I was interested in your report of Michel Valstar's work on an algorithm that could help spot autism-like conditions from facial expressions (7 January, p 11). Many autistic individuals find our eyes drawn to the mouth rather than the eyes of others we're conversing with. Could this be used by the researchers?
Michel Valstar writes:
• This thought resonates with our work with Alexander Foss, an ophthalmologist interested in the development of autism-like conditions. Gaze analysis is one cue we will include in future work.
For the record
• Our piece on the work of Dora Colebrook and Leonard Colebrook on streptococcal infections appeared with a photo of someone else (4 February, p 40). See the online version (bit.ly/Colebrook).
• Matthew Hodson of UK charity told us that studies of couples in monogamous relationships, in which one partner is HIV-negative and the other positive and on treatment, mean: “When we are undetectable, we are uninfectious. This means that pretty much all the fear that HIV-negative people have of those of us living with HIV is just wasted energy” (11 February, p 22).
• Oh caption! My caption! The first written language emerged some 5400 years ago, as the main text of the article stated (11 February, p 34).