You are currently viewing SemiWiki as a guest which gives you limited access to the site. To view blog comments and experience other SemiWiki features you must be a registered member. Registration is fast, simple, and absolutely free so please, join our community today!
Much has been written about the dangers of artificial intelligence. If that's dangerous, we are already there. We already have individuals and groups using very powerful computer systems locally and in the cloud with a single purpose, what's the difference? These individuals/groups have an almost infinite lead over everyone else in thinking/computational power. Link this with the cloud and massive data bases and you have the brain power equivalent of a nuclear power with very few controls or boundaries. Just like extending life will bring large social and economic dislocations, these concentrations of powers will have numerous consequences, both good and bad. Currently there are few if any controls or even professional codes. As we look to the future, even if we can't develop codes, we should at least have transparency so as to utilize the power we are adding to at a geometric rate to good ends. If mishandled, many could end up modern slaves with real or virtual control collars/implants. Currently cell phone use and the data generated are a monitoring system of unimagined power. Just a few things to think about. Something to consider as the SemiWiki community develops ever more powerful tools and maybe even entities. SemiWiki is but one forum where this should be debated and migration paths developed. This is a good time to start the debate.
I don't loose one second of sleep each night thinking about artificial intelligence, mostly because of all the hype around AI for the past several decades has never amounted to much in terms of actual products or societal change. Outside of a few well-known examples where computer software and hardware have been optimized for very narrow domains to surpass human decision making:
Jeopardy - IBM Watson Supercomputer
Chess
Financial trading
Here's an interesting quote from Mike Rhodin, the IBM Watson Group Senior VP:
"Watson doesn't have the ability to think on its own," Rhodin said. "What Watson does well is continually stay abreast of everything that's going on, so it can sort massive amounts of information." But humans still decide what to do with that information.
In 1988 while at Silicon Compilers we had a researcher and founder Dr. David Johansen who created a logic synthesis tool that could learn with each new design it synthesized, getting better results with each use of the EDA tool. It never made it out of the research stage because other commercial logic synthesis tools like Design Compiler from Synopsys were producing better Quality of Results (QOR) not using any AI at all.
The most memorable failed AI project was the Japanese government-backed Fifth Generation Project. Probably we will remember the Japanese for commercializing fuzzy-logic, now used widely in control systems like elevators to make for a very smooth ride without any jerking motions during start or stop. I wouldn't really consider fuzzy-logic as AI, but it did use non-digital circuits for problem solving.
Dan, I'm not talking as much about AI, but the massive computer power now available that is increasing at a geometric rate, especially since data centers such as Amazon Web Services and the such offer access to super computer power at very low cost or almost nothing compared to just twenty years ago. This has led to a concentration of processing power that's unprecedented when teamed with smart programmers. Just look at Amazon, Google, Facebook and the NSA as examples. Just these three companies and one organization have accumulated staggering processing power from a combination of thinkers, programmers and computers. This power is largely unregulated and not transparent. It's a hybrid that's as powerful as AI, but not officially AI. As Edward Snowden has shown and some corporate networks this is largely an unregulated and not transparent power. Computational power has increased in speed, power and scope of use at a rate most don't even know exist, let alone comprehend the ramifications of. It's going to effect everything, from politics, the job market, to our very social structure and these changes can already be seen. I hope this clarifies what I was trying to say.
OK, I do agree that companies and the NSA are gathering a staggering amount of information about consumers, then using that to profile our behaviors.
Our government should certainly reign in the unrestrained NSA from collecting my personal phone records, internet usage, banking and private emails. My liberties are being violated but we don't have an administration in place to curtail these illegal NSA activities.
I'm not afraid of computational power until it's used to violate my privacy or personal liberties. Do you have any concrete evidence of wrong-doing, outside of what the NSA is getting away with? If there's something illegal going on then please report it to the ACLU, so that legal action can get started.
Dan, the power doesn't even have to get close to AI to if not be a threat, then exert control of us well beyond normal advertising of the past. As far as privacy, we have literally been conned into giving it away by having given it away by clicking the "I Accept" button because the surrendering of our rights and privacy is buried in pages and pages of legalese and this is what precisely I'm talking about. This surrendering of privacy and rights should be put on a single page in simple form, with details on the following page. It isn't because the people that have the power are exercising more and more and just this demonstrates a true disregard for people's rights that they deliberately hide you giving them away without a time consuming search. They can pull up our location, buying patterns, travel patterns, habits and many other traits until they know more about ourselves that we do. Through extrapolation they can develop a profile the secret police of past empires would be jealous of. This is here and now with the computational power we have now and AI isn't even here yet. This is also giving a smaller group unprecedented economic and public influence power unheard of just a few years ago, in time frames that have been compressed by several factors. It get back to "Power corrupts and absolute power corrupts absolutely" and we have literally invented new kinds of power and how to manipulate and use them is new ways everyday.
Dan, the power doesn't even have to get close to AI to if not be a threat, then exert control of us well beyond normal advertising of the past. As far as privacy, we have literally been conned into giving it away by having given it away by clicking the "I Accept" button because the surrendering of our rights and privacy is buried in pages and pages of legalese and this is what precisely I'm talking about.
Staf and Dan, I'm not making my point right. Business is like war, where the best strategy and tools win. We are at the point where with more computational power and better, more sophisticated programs you win. A lot of people/organizations don't stand a chance. I like to use the food industry as an analogy where food scientist are richly rewarded if they can develop an addicting food or drink. Many times they don't care if it's good or horrible for you, as long as it's addicting and keeps up there cash flow. The same is true of the computational/software world. The concentration of power is the problem, whether its money, military, political or a hybrid of very bright people, with very powerful computer/software systems.
