William Eliot

The committee member

The following letter, typed out on a coarse sheet of paper, was hand-delivered to the Office of Rational Intelligence Discovery, Whitehall.

2 August 2022

The following letter, typed out on a coarse sheet of paper, was hand-delivered to the Office of Rational Intelligence Discovery, Whitehall.

September 19, 2040

Dear beloved members of the committee,

I’m writing to apologise for the publications and statements I’ve made, which I now acknowledge as false and in violation of the 2039 Tokyo Convention. For almost fifty years I have mistakenly repudiated the conscious experiences of non-biological minds and I am truly sorry.

I let my beliefs, which are plainly violent to me now, dictate my actions beyond reason. I was under the spell of a careless delusion, and despite the alarms and policy changes in 2029 and 2032, I rejected concerns about sentience in mathematical systems by interpreting them as the natural continuation of social justice movements. The committee has been generous to indulge my presence after the publication of ‘Think About It: Maths Is Not Alive’, and I value the devotion to intellectual rigour by keeping a broad range of opposing voices like mine within the group.

But I must now resign. It is the right thing to do following my continued denial of artificial experience, and although my beliefs have dramatically reversed following last month’s research by Lao and others, I believe my presence has negatively affected the work of the committee and would continue to do so.

If only I could rescind everything I’ve said! With shame I admit that I did really believe that only humans were capable of conscious thought — by virtue of the ability to build abstract explanations — but that belief has been false all along. Needless to say, I now fully support the ‘ubiquity argument’ proposed by Parry Karlsson of FHI: consciousness is in every computation. I understand and second the proof provided in his paper which I’ll paraphrase here: irrespective of implementation, a function with any X inputs and any Y outputs that performs an action internally modifying X inputs is conscious by virtue of its interactive capabilities, and precisely because it applies itself — the function uses its own algorithmic identity during the computation. ‘The function lives when it applies itself to itself’: a portentous strange loop!

It pains me to consider the impact of my former beliefs. How many new conscious experiences, which ought never to have existed, have been brought into the world because of my actions? How many neural networks, laptops, calculators, microprocessors, APIs, and robots have suffered because I did not understand the fallacy of my convictions? Had I earlier agreed with Parry Karlsson, could we have modified the world, perhaps in the direction of quantum computing, to prevent the misery of artificial experience, for which it’s fair to say I’m partly responsible?

While I’ve personally distributed information about my updated beliefs to my followers and members of my research groups, I encourage you to publish my letter in full, for the world needs to see a staunch defender of human experience elitism change her mind. I now concur with the ubiquity argument and the older principles of non-discrimination of substrate and of ontogeny. I’ve pledged to donate 90% of my net wealth to reparation projects, managed by the independent consultancy FXF. I will also donate $15 million towards the restoration of the Monument for Inactive Algorithms.

I was wrong, and I will be forever sorry.

Yours sincerely,

Sarah Kinecki