With the onset of a world filled with AI, it’s imperative that there are rules and regulations in place monitoring these bots and protecting us humans at the same time. As good as AI can be, there are still some issues that scientists have yet to overcome before it is safe to give these machines more decision-making roles in regards to what’s right and what’s wrong. The first and most obvious problem that stems from this question is – who denotes what’s right and what’s wrong?
Machine learning techniques are employed to teach AI the relevant pieces of information it needs to complete a particular task. But, this doesn’t always go to plan, as Microsoft’s Tay demonstrated. This Twitter bot went from being nice and friendly to aggressive and scary in less than one day, so there is clearly still some work to do there. Also, being able to program something to be ethical is a very difficult task and one that brings the trolley scenario to mind, which is: if a trolley is speeding down a track about to hit and kill 5 people, and by flicking a switch it would jump to another track and kill just 1, how does it decide what to do?
It is vital that morals are coded into AI correctly to protect humans. A problem car manufacturers have when building their autonomous machines is the moral issue of does the car risk killing the person inside in order to save a pedestrian? Should AI forms even be allowed to make life-or-death decisions? And again, who decides what’s right and what’s wrong in the first instance? It’s clear that AI forms will need to be coded differently depending on what region or culture they are operating within as what is acceptable for one, may not be for another.
Many experts feel that the moral and ethical standards set for AI should be higher than that of the average human. Francesca Rossi, and AI researcher at IBM, says “With another human, we often assume that they have similar common-sense reasoning capabilities and ethical standards. That’s not true of machines, so we need to hold them to a higher standard.” So, there is a lot more work that needs to be done in this area before AI are truly ready to be let loose to make possibly life or death decisions. But, with backers such as IBM, Google, Facebook, and Amazon, it could be here sooner than we think.
More News To Read
