1
$\begingroup$

Is artificial intelligence and, in particular, neural networks being used in real-world critical applications and devices?

I had a discussion with my colleague who states that nobody would use artificial intelligence, especially neural nets, for critical stuff, like technical devices or sensors.

I'm only aware of the problem of neural nets being so-called black-boxes, but, nevertheless, I think it is possible to make an NN robust so that it matches the demands of daily processes, also in sensitive fields like health care, energy market, self-driving cars, and so on. Yet I cannot underline this.

Does somebody have more insights or other information, opinions and so on? I appreciate any meaningful answer.

$\endgroup$
  • 2
    $\begingroup$ Your definition of critical stuff seems a little vague to me. $\endgroup$ – DuttaA Oct 17 '19 at 14:38
  • $\begingroup$ @DuttaA Sure, let's say something where the output of a sensor can lead to huge damages (in whatever way). $\endgroup$ – Ben Oct 18 '19 at 5:17
1
$\begingroup$

This is as much an ethical concern as a practical one.

AI systems are already reaching or exceeding human performance in many critical areas. Consider the detection of common cancers, where AI systems match or exceed humans. Another good example is Tesla's Autopilot, which is actually safer than human drivers, but gets a lot of bad press when it makes a mistake. Both of these systems are likely using Neural Networks, possibly alongside other heuristic or rule-drive approaches.

The issue isn't whether these systems can be "safe enough" for everyday use. They are safe enough in a societal sense already. The concern is that the people who die when these systems make mistakes are randomly selected, whereas the people who die when a human performs the work die because a human makes a mistake (usually). This is difficult to accept for the same reason that some people are scared of flying, even though it is many times safer than driving the same distance: there is a loss of control, and "good" people may die through no fault of their own, or perhaps through events that are no one's fault.

Whether we use this systems will thus probably depend on the application. In Medicine it's easy to see a case: we can do more tests than before for the same price. People who can afford to have a doctor review the machine's decisions are probably no worse off. People who couldn't afford this already are better off (they get a diagnosis with some positive predictive power value now, instead of none at all). In driving, it's more complicated, and will probably require further development. No one knows for sure how good self-driving cars can get, but they'll probably get somewhat better over time. Maybe they'll get good enough that they more or less never kill people, or maybe they won't.

I actually think your friend is way off about the use cases he thinks AI is not suitable for though. Check out this article in Military Embedded Systems. In applications where decisions are made on a pure cost/benefit basis (and not based on people's gut feelings about morality), AI systems are actually easier to adopt, and are already often better than human operators. This trend seems likely to increase in the future, so "technical devices and sensors", which are often black-boxes to the lay public anyway, seem like they are among the first things to go.

| improve this answer | |
$\endgroup$
  • $\begingroup$ You should be careful with "People who couldn't afford this already are better off (they get some diagnosis now, instead of none)". It may be the case that no diagnosis is better than a wrong diagnosis, but I understand your point: you want to say that there is more opportunity. $\endgroup$ – nbro Oct 18 '19 at 0:56
  • $\begingroup$ @nbro Good point, I'll change the wording. $\endgroup$ – John Doucette Oct 18 '19 at 2:09
  • $\begingroup$ @JohnDoucette Thank you. The linked articles are showing promising but yet ongoing researches/developments. I would like to know if there are NNs deployed to the real world (besides speech recognition or chatbot and stuff like this - which can be criticial but shouldn't be primarily). $\endgroup$ – Ben Oct 18 '19 at 5:29
  • 1
    $\begingroup$ @Ben A good example is IBM's Watson for Oncology (ibm.com/us-en/marketplace/clinical-decision-support-oncology). You can buy this product today, and it is already in use at various hospitals. Tesla products are also something you can buy today. There are lots of companies like this, but they're usually focused on B2B sales. Here's a list of a hundred AI startups currently selling in the medical domain: beckershospitalreview.com/lists/… $\endgroup$ – John Doucette Oct 18 '19 at 13:34
  • 1
    $\begingroup$ @Ben Telsa is not driving fully autonomously, but Tesla's autopilot already appears to be safer than human drivers for highway driving, which seems critical enough (see the story in my answer). Waymo (techcrunch.com/2019/10/09/…) announced a level 4 autonomous vehicle will be available for ride-hailing in Phoenix later this month. You can also check out products like Buoy (buoyhealth.com/employers), which are for sale and use AI to perform medical diagnostics. $\endgroup$ – John Doucette Oct 18 '19 at 13:43
1
$\begingroup$

There is significant lag between overseas military operations and people operators back in the states. Because of this, the military is offloading more functionality to the remote piloted drones and convoys. The same will be the case once we have functional rescue bots. Look up boston robotics, if you don't think robots will be entering fires or ravines in the near future.

| improve this answer | |
$\endgroup$
  • $\begingroup$ I thought about the Boston robots, too, but, I would say when they come into play they can hardly "destroy" anything as, in case of military this their purpose and in case of firefighting, this will be the case without them anyway :) $\endgroup$ – Ben Oct 18 '19 at 7:48

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.