Transformative Impacts

Two years ago this month, CRN was founded. Our mission, as formulated then, was "to raise awareness of the issues presented by nanotechnology: the benefits and dangers, and the possibilities for responsible use." We remain committed to those ideals, although our activities are now moving beyond raising awareness and toward engaging and empowering many groups to prepare for the transformative and potentially disruptive impacts of molecular manufacturing.

Of course, nanotechnology is not the only field in which scientific progress and technological applications will hugely affect society. Biotechnology and genetic engineering, space exploration and development, surveillance and privacy issues -- these are just a few of the others (each of which, by the way, is also entwined with nanotechnology).

If one is paying attention, the rate of change can seem dizzying. We're now even hearing about the real possibility of bloodstream robots, for example. Also, it's clear to many people that serious attention should be paid to the consequences -- both good and bad -- of developing artificial general intelligence, sometimes known as Strong AI.

In the current issue of Legal Affairs, Benjamin Soskis offers a fascinating essay about the thorny issue of how we might grant legal rights to computers. Here's how it begins:

Last year, at a mock trial held during the biennial convention of the International Bar Association in San Francisco, Martine Rothblatt argued an especially tough case. The difficulty for Rothblatt, an attorney-entrepreneur and pioneer in the satellite communications industry, was not that she represented an unsympathetic client. Far from it—the plaintiff's story of confronting corporate oppressors moved the large audience. The problem was that the plaintiff was a computer.

According to the trial scenario, a fictitious company created a powerful computer, BINA48, to serve as a stand-alone customer relations department, replacing scores of human 1-800 telephone operators. Equipped with the processing speed and the memory capacity of 1,000 brains, the computer was designed with the ability to think autonomously and with the emotional intelligence necessary to communicate and empathize with addled callers.

By scanning confidential memos, BINA48 learned that the company planned to shut it down and use its parts to build a new model. So it sent a plaintive e-mail to local lawyers, ending with the stirring plea, "Please agree to be my counsel and save my life. I love every day that I live. I enjoy wonderful sensations by traveling throughout the World Wide Web. I need your help!" The computer offered to pay them with money it had raised while moonlighting as an Internet researcher.

Far-fetched? A waste of time? Or, perhaps, a worthwhile exercise in preparing us to confront one of the unprecedented challenges of the near future.

Mike Treder

Comments

[Brett Bellmore]

"But the mock trial judge, played by a local lawyer who is an expert in mental health law, set aside the jury verdict and recommended letting the issue be resolved by the hypothetical legislature."

Now, THAT is a disturbing prospect: You sue for your right to not be killed, the jury rules in your favor, and the judge sets aside the verdict, allowing you to be dissected.

An important point here, which Soskis missed, is that, given the short design cycles, and easier modification, the age of human equivalent AIs is likely to be just a moment, before they race beyond us, leaving us in the dust. So when that moment comes, perhaps we want to be a bit open minded, grant them the benefit of the doubt... Because the question will shortly be, will they grant US rights? And it would be nice if they approached that question with memories of our good will.

[Brett Bellmore]

In fact, I suppose it's entirely possible that we'll transition from subhuman to superhuman AIs, without any intermediate step. After all, by the time we understand intelligence well enough to engineer a genuinely general purpose, human level AI, we'll be capable of building hardware much more powerful than a human brain. And the utility of something that can do things people aren't capable of is orders of magnitude greater than something that can do things within human reach, but merely cheaper.

[michael vassar]

This is not a realistic scenario. Any AI capable of acting human enough that people other than science fiction fans would consider granting it rights would certainly be capable of defending itself, unless it was a very slow or otherwise limited upload, in which case the original person from which it was produced would have rights and it would in all probability not differ from that person enough to sensibly consider it to have diverged into seperate personhood.

[michael vassar]

It it pure anthropormophic fantasy to imagine that the decisions of AIs as to whether to grant us rights would be based on reciprocity. That could happen if their motivational drives supported reciprocity in a manner which was robust to self-transformation or in a manner which was supported by their environment, but if we could guarantee that we could guarantee benevolence without the need for reciprocity.

[Brett Bellmore]

Not necessarilly; We could get AIs by "porting" human neural patterns, either generic or specific, to artifical media, in which case they'd be just like humans: Hardwired for reciprocity, but benevolent only if treated well.

[Michael Vassar]

I mentioned the specific human neural patterns possibility in the uploading caveat, but pointed out that only a very limited or slow upload is relevant, as a powerful one can self-modify recursively and human reciprocity wasn't built to be stable under recursive self-modification (which I called self-transformation). I suppose that a person with no ability or inclination to self-modify could be uploaded, for a while, but in the long term the restriction of uploading to a poorly defined class of rapidly diverging individual will not hold up for long enough to make rights issues important.
Porting of generic human neural patterns is almost certainly the easiest way to get human equivalent AI, but is probably a very bad idea, as raising such an AI well would be terribly difficult in the absence of a very powerful technological base, and the time required before such an AI would be useful would probably be prohibitive for most economic applications. Trying to speed up development would be asking for trouble. I suppose that with MNT to make bodies, unethical people might use mass produced electronic four year old wireheads in robotic bodies as inexpensive slaves. Mixing human and animal neural patterns might enable faster maturation to useful age, though this is also asking for trouble. These are trueyl Ugly thoughts though. Lets try not to build that kind of world.

[Brett Bellmore]

I agree that some of these concepts are quite ugly; That's one of the reasons I'm dubious about the notion of artificially intelligent servants of human equivalent, let alone greater, intelligence. From a social standpoint, we'd be much better off pursuing a stratety of intelligence enhancement, mating improved problem solving engines to existing human minds. Extending our existing brains the way the frontal cortex extends the brains of our primative ancestors.

[Todd]

hello all :)

sorry i have been away for a time

i am looking for a new verbal imput software

the ver 4 dragon i have will not work on my new XP computer

anyone know were i can download something for free or cheeeeeep

thanks and i look forword to long thoughts on many things great and small :)

todd

[michael vassar]

Trouble is Brett, we can't just ignore a tech cause it would be ugly, we have to figure out what it would do, who would make it, etc.

[Tom Craver]

Todd - I'm assuming you have a disability that makes typing difficult or impossible? I don't know where you live, but here in AZ, my daughter was provided a wheelchair and other equipment she needed. I would presume that if you checked out local/state assistance groups, there may be some that would be happy to provide what you need? If you're in the US, you might google with your state's name and "assistive technology" - that seems to work well.

[todd]

SYLLABICATION: rec·i·proc·i·ty
PRONUNCIATION: rs-prs-t
NOUN: Inflected forms: pl. rec·i·proc·i·ties
1. A reciprocal condition or relationship. 2. A mutual or cooperative interchange of favors or privileges, especially the exchange of rights or privileges of trade between nations

I looked this one up as it is a very interesting word :)

[todd]

Well, I got a copy of dragon seven at Staples and on sale today, so I picked it up and am now ready to continuing comments in regards to AI.

The artificial intelligence discussion has been long-running and argumentative over the issue of feasibility. This fundamental argument seems to be falling to the wayside as those who believe in the technical feasibility of strong AI began to outnumber those that do not believe. The above article brings up an interesting point one I had not considered in general when I think of strong AI I immediately think of robotics and the impact unlimited manpower will have on the society through the use of robotics. But a second and perhaps more immediate concern would be the impact AI would have on industries which utilize over the phone or helpdesk situations.

In these industries for example 411 operators a strong AI replacement could be implemented and running in a relatively short period of time compared to the design and manufacture of full-scale robots capable of performing real work. A fue years back I worked a contract where I replaced computers for Pacific Bell we replaced some 25,000 systems or the course of six months all of the systems were used by operators considering multiple shifts utilizing one computer one could assume some 50,000 individuals plus were employed with this one company utilizing one aspect of the helpdesk industry.

As undoubtedly many books and articles written on AI I would be interested in anyone's opinion as to the general timeline once strong AI is present as to its impact on specific industries and eventual impact on all industry.

It is my opinion that industries utilizing over the phone and data related industries would be the first to fall or be critically undermined to the end of showing a 80 plus percent reduction in manpower for a given company.

Following this line of thought given the existence of strong AI on day one what is the impact on day two and those days to follow? To all companies currently existing in United States and trading throughout the world. It is in my opinion easy to describe situations where almost any scenario is plausible at strong AI plus five years. But a more interesting discussion would be carried through when discussing AI plus one day.

Several writers here have stressed the possibility that AI will in some way wish to conflict with the current path. If we were to simply try to move forward and “assume” these individuals are mistaken in their assessment that strong AI will in some way conflict with our current lifestyle. We can then begin to address what changes will occur or perhaps could occur given the existence of strong AI. The impact on the helpdesk industry is certainly profound and should be noted. Additionally assuming the construction of robotics utilizing strong AI this also represents great change and a deliberate impact on manufacturing and other fields.

I would be curious to know what additional areas and industries are being discussed for impact when strong AI becomes available.