2025-02-04
It does seem that we live in a time of great peril to the ordinary man or woman on the Clapham Omnibus.
AI is on the threshold of becoming ubiquitous in many spheres, but in none more so than in matters of government.
The reason of course is that AI tends to be big complex and expensive, and especially reliant on Big Data, and Big Training. It is also largely beyond the reach of those without Big Wallets Big connections and the aforementioned Big Data.
But it might bring the Big Statute Book within reach of Joe Citizen for specific queries.
So the main customers for AI will be Big Governments and their associated corporate contractors who will provide the interface between us and them.
It is also currently not so easy to comprehend the veracity of its results for any given sphere of enquiry, whilst being well within the means of the usual corporate suspects to design and deploy - and their customer will not be "us", it will be "them".
Should we trust it? How can we trust it? Is it even practicable to challenge it1? Do we have any protection at all?
1 The EU has begun the process of legislating for it. But unless every AI can, together with its results, provide a supporting list of assumptions and facts upon which it relied in order to come to its conclusions, then the chances of challenging them remain miniscule.
