2 ms·
The (well known) problematic bit about AI is that it furthers the diffusion of responsibility that already became so commonplace with computers. People can jus
by atoav 2y ago
The (well known) problematic bit about AI is that it furthers the diffusion of responsibility that already became so commonplace with computers.
People can just handwave catastrophic decisions away with a "the computer made an error, nothing we can do". This has been the case before AI, the differrnce AI makes is just that more decisions are going to be affected by this.
What we need is to make the (legal) buck stop somewhere, ideally in a place that can positively change things. If you’re a civil engineer and your bridge collapses, because you fucked up the material selection, you go to jail. If you are a software engineer and you make design decisions in 2025 that would have had severe security implications in the 80s — and then this leads to the leaking of millions of medical records you can still YOLO it off somehow and go to work on the next thing.
The buck has to stop somewhere with software and it doesn't really. I know that is a feature for a certain type of person, but it actively makes the world worse.
- hkwerf 2y agoThat is explicitly addressed by the regulation. If an AI product is used in a way you describe it, there need to be measures put in place to reduce the risk to an acceptable level. If these measures are missing, it cannot even be put into service. This places the responsibility for the use of AI clearly with either the entity that places the product on the market (i.e. if you sell something in the EU) or that puts it into service (if a EU company uses an AI product across borders remotely). With the required employee AI training, this is further shifted onto the users if done correctly or stays with the employer if not.