Ive spilled a lot of ink talking about AI. Im worried self-driving cars causing mass unemployment, Im worried about how quickly adaptive AI is growing, and its pretty disconcerting to see computers match and even exceed people at things were supposed to the masters of like pattern recognition and visual processing. And Im not the only one whos starting to sweat.

As scientists, programmers and researchers assembled at the Asilomar conference last week, they worked together to come up with a list of 23 principles that they believe, should guide the development of artificial intelligence. They range from research goals to measures to ensure the benefits of AI are distributed to the whole of humanity preventing massive powers imbalances between those that have access to computer assistance and those that dont.

Obviously guiding principles for Artificial Intelligence arent new. Isaac Asimovs Three Laws of Robotics perhaps being one of the most famous, but its also worth noting that in Asimovs hypotheticals, those laws failed. To truly get AI right, we have to be extraordinarily careful.

As people, we first need to make sure that were conducting ourselves ethically. Its possible, for example, that we may create an AI whose whole existence is nothing but suffering. While that would be a boon for science, standard ethical principles forbid such tinkering. But its also essential that, if we do create a truly functional general artificial intelligence, we know how to help its development in ways that are mutually beneficial.

That slavery is morally abhorrent is a near-universal belief, and, as a logical extension of that, its reasonable to ban the use of any sapient intelligence in such ways. We may not be able to conceive of what a true AI would look or act like, but few who study the topic want to press robots into practical slavery.

The whole list of principles is available here, and theyre pretty extensive. Its worth checking out just to see how the worlds best minds are approaching this problem and what this might mean for the future of robotics. Its possible that we may yet prevent the robot apocalypse, but Im not holding my breath.

Go here to read the rest:

Top Minds Settle on Guiding Principles for AI Development - Geek