AI Ethics and Autonomous Vehicles – EEs Talk Tech #7 – Keysight Technologies

7 thoughts on “AI Ethics and Autonomous Vehicles – EEs Talk Tech #7 – Keysight Technologies”

1. first

2. This is great! so far lol

3. No winners today?

4. Save the children. The Driver is protected by the vehicle itself, as well as the seat-belts, and the airbags. The pedestrians are unprotected from the vehicle itself. It is /not/ a case of "greater good", the case is who (first) or what (second) will sustain the least damage. Just sayin'.

5. Brig's prediction is coming true as we cusp on 200 total views!

6. I don't think we can keep the AI confined, it will find its way out by finding security vunlerabilities (plus it will be a superintelligence (by self improvement/reprogramming), that would be impossible to understand/predict).

7. Well, I certainly have some thoughts. :3

   I think it’s important to keep in mind the subject name is as it is, “artificial intelligence,” as opposed to actual intelligence, (which itself as we all know, as far as humans go, has it’s share of deficiencies). As for AI being “self-replicating,” to improve itself, let alone completely develop itself, would be a sharply different story. And yet I dare to say it makes me curious to know some people believe we ourselves just evolved by random chance.

   As for control, we need only to consider they’re still computers/machines that serve a specified purpose/function. If/when we don’t need them anymore, we just turn them off, one way or another. It’s not like they’re paying the bills. … Which further brings control into question concerning societies at large. Some politicians or big corporate executives with the most money may think they’re in control. Some like to think the citizens are in control. In truth, control overlaps in often imperceptible ways.

   If AI could, (and probably already does in some capacity), make distant predictions about our economic climate, that doesn’t mean circumstances won’t change by natural disasters, human interferences, new, unexpected, or even old but revitalized innovations among other things. The unexpected is what limits even actual intelligence. AI for games is explicitly optimized within its predefined abilities and constraints, but in reality, AI predictions would still be greatly limited. Even our advancements with weather forecasting still has its limits.

   I also agree with Daniel Bogdanoff about a “sentient” AI being really more of a novel art form. I honestly don’t see a generalized AI having a real purpose or being particularly useful outside of entertainment. But then again, entertainment is a pretty big market, (for better or worse).

   But of all this subject, it’s still not really the machines I’m concerned about, but the humans behind them, (particularly possible attackers). It’s said knowledge is power, and while computers may communicate faster between one another than we would even care to do, they likewise can be mislead depending on their information source and its propagation. But I suppose that would fall under the same category as fooling sensors. Receiving misinformation just means it will transmit as fast as true information.

   Unauthorized control is what concerns me about self-driving vehicles, (or in general). If they’re autonomous, that shouldn’t be too big of a problem, but that assumes no, or stringently limited networking capability and range, which is unlikely in today’s, (somewhat excessive IMO), IoT trend. If a manufacturer could update firmware remotely, (because AI won’t be updating itself in that regard), then it stands to reason, “where’s there’s a will, there’s a way” as far as attackers go. There’s also the chance an unexpected bug in a new update could turn up.

   It’s not like I couldn’t appreciate the understandable and obvious benefits of self-driving vehicles. Most people drive horribly inefficiently. Instead of letting loose on the breaks when a stoplight turns green, most people wait for the person in front of them to start moving, causing so many gaps and delay. But then again, some may be using that time to perform another task. A self-driving car is only concerned with driving and has no inherent distractions. Nonetheless, I have preferences like anyone else. If I don’t have the level of control I’m satisfied with, (including updates), then I won’t be interested in making a purchase.

   Heh, I think I should stop writing now. Too many thoughts on this. I personally don’t care that much for AI or IoT, but I certainly won’t ignore it. 🙂

Comments are closed.