We are confronted with so many warning labels that they no longer register when we see them. This is one that should catch everyone’s attention. Any artificial intelligence technology that presents a risk to the public should come with clearly written warnings and disclosures in order to allow the public to make informed decisions when purchasing or even being around this technology. Buildings in California already display warnings about carcinogens, should highway warning signs for autonomous vehicles be next?
Product disclosures are a challenge the artificial technology industry must address. If you are buying an autonomous vehicle, you are going to want to know what the exact algorithm is for deciding when it will likely injure or kill you over someone in another vehicle or pedestrians on the sidewalk. Or better yet, you may want it to always decide in your favor. That should be a simple enough programming change, but there is a good chance that manufacturers will make it a feature available on only the most expensive models. Will the preferential programming of devices be for sale only to those who can afford it? Can our institutions create and enforce artificial intelligence device equity within our society that has such a large income and wealth equality gap?
Giving the consumer a choice of programming is a challenge in itself. The progress of machine learning has already begun to outpace the ability of programmers to understand the code being generated. This makes customization difficult at best and perhaps impossible as the quantity of code and level of complexity accelerate. How can the relevant risks for the consumer be identified and disclosed if the exact mechanisms involved cannot be understood?
The regulation of technology employing artificial intelligence is in it’s infancy, as is the technology. Governments and other institutions, including academia, are poorly equipped to provide rigorous and independent evaluation of the impact these new technologies will have on our society. Due to the huge amount of investment money flowing into the field, expertise in artificial intelligence is being monopolized by industry leaving fewer resources for regulatory efforts. The absence of meaningful oversight creates risks for society and those who stand to generate enormous wealth from these new technologies.
Autonomous Vehicles will prevent many deaths and injuries in numbers far greater than any that may be caused by the adoption of this technology. There is a moral obligation for our society to exploit this technology. Undoubtably, there will be many bumps on the road to success, but the benefits in human terms are far to great for us to be deterred by technical or social challenges. Here is one final warning on the subject.