Another time – The Economist newsletter, April 28th 2023. Subject line:
Yuval Noah Harari: regulate AI before it regulates us
Such headlines and statements make me think of the phrase sweet nothings and I am wondering if in political-academic respect a similar phrase must be formulated: wise nothings, not to say nothing wise.
The first point – formulated as question:
Is it really about regulating AI or is it about
regulating the making of AI
regulating the use of AI
regulating the use of what AI “does”
regulating the – and then: which – users of AI
?
The second point – put forward as kind of dystopian saga, and made concrete, real – seems to be more important.
We witness major progress in computer technology. The quasi-ancient times of the use of computers required somewhat simplified communication with the figure zero and one – in other words: while not speaking directly in sequences 0 1, the programming languages was not so different, just slightly: presenting the binary code in a form that looked – in part – like real language. Today, computers can be used, directly employing real language – and even some sloppiness will be accepted: different pronunciation, spelling, terms … are no problem for the processing.
The next step is the exciting one: just one word, term, short question or statement translates into kind of explorative story. Specific version: upload a document, e.g. an edited volume on racism – the stunning result: you key in an issue you want to look at and you will receive a perfect summary, completely informed and elaborating on the contradictions of opinions expressed in the different contributions. Then we have a machine translation, at this stage working reasonably good. On top of all this we have now a language correcting software, translating miserable English into good English, awkward French language into well-formulated French … . The problem is only the programming of well-functioning links between different AI instruments.
Just key in a term, perhaps name… . The result will be posted automatically to the translator, then to the AI proof reader and you will get an article on something you never heard about, written, even thought in the style of a person you don’t know in a language of which you never heard, spoken in a country of which you do not know – authored by yourself. In other words, you will not know anymore what you are doing – however, in this case – and looking at the German Criminal Law – one is wondering if we are facing a reformulation of paragraph 20, now reading
Whoever, at the time of the commission of the offence, is incapable of appreciating the unlawfulness of their actions or of acting in accordance with any such appreciation due to a pathological mental disorder, a profound disturbance of consciousness or intellectual disability or any other serious mental disorder is deemed to act without guilt.
A new version must then contain an additional sentence, like
Non-existence will not be considered as excuse, qualifying people as acting without guilt.
So, it remains your decision: feel yourself as genius or as basically non-existing something. You may translate this into “control AI or control yourself”.
COMMEDIA DELLA VITA OR PÁNTA RÊI'S FIRM GROUND 