IBM Unveils System That ‘Debates’ With Humans

How to destroy all AI machines: Time flies like an arrow, but fruit flies like a banana. Why?

As with all technology (see Facebook, for example), AI will end up stunningly useful in small, compartmentalized arenas, and simultaneously completely pollute general humanity. Thus when we all have AI robots, they will spend all day sitting on our couches looking at porn through our VR headsets.

It's worth a critical look at how disappointing Watson has been in health care before we declare this newest initiative the next step in machine evolution.

Dr. Frank Poole: Well, whaddya think? Dave Bowman: I'm not sure, what do you think? Dr. Frank Poole: I've got a bad feeling about him. Dave Bowman: You do? Dr. Frank Poole: Yeah, definitely. Don't you? Dave Bowman: [sighs] I don't know; I think so. You know of course though he's right about the 9000 series having a perfect operational record. They do. Dr. Frank Poole: Unfortunately that sounds a little like famous last words.

Intelligence requires *mind* Naming this field of engineering "artificial intelligence" is either delusional or monumentally optimistic and hubristic.

There is a discrete between an output of a computer replying to the programmed set of questions & answers based on the established facts of STEM (Science,Technology,Engineering & Math) and the Cognitive analysis in the CPU and guide / lead personnel to the next step forward or withdraw. We can't insert any information beyond the array of pertinent -programmed facts based on experience and experiments - i.e simply a podcast of Wikipedia or the preset sequence.The basis for my argument is the root cause analysis of the first victim to the driverless car in California. Barring a capability for a cognitive analysis of its own the AI & Robotics can't create a decision maker who could compete with a human being. So far we are successful with a transmitter and a transducer but we haven't reached the stage of creating a robotic-decision maker that could decide of its own to perform a transmitter or a transducer in the new environment with reflex. AI would become a trendsetter when it has reflex to the unexpected event in the environment.

So, IBM could make an AI system that would be able to become POTUS, as it could easily win a debate with Trump.

The important thing is to make these machines transparent so we know why the machines are doing what they do so we understand what they are doing what they are doing. It is very important that the developers of AI determine how these machines work so we know if there are any prejudices or preconceived notions in their programming. As they are developed by humans it would be surprising if these machines are prejudice free. Until we include transparency in the development of AI it will be difficult to believe advanced interactions with AI computers.

Can we put two of these machines to debate each other, one for and one against human existence? Their arguments will be studied for years to come.

For 200 years science fiction writers have been predicting fantastical technologies. Many of them have already come to pass. They have also been predicting disasters that come from these technologies. Many of these disasters have come to pass. Science fiction writers see a lot of promise and danger in artificial intelligence. They have proven to be prescient. We should heed their warnings and be careful about how we implement these technologies. Many say, "people are always afraid of the new, but we have to keep moving forward anyway." That may be true, but that doesn't mean that we haven't almost had nuclear Armageddon, or couldn't tomorrow. We have to think. We have to debate. We have to make sure the economy is designed around people not machines. Capitalism is all about capital, i.e machines. Democracy is about humans. We have too make sure that democracy can stay one step ahead of artificial intelligence which someday will be able to manipulate us even better than Trump does. Artificial intelligence will not necessarily be obvious when it suddenly works. Code is already writing code that we don't understand. When AI suddenly works well and is smarter than a human, it may forget to mention it to us. I'm not saying don't create A.I., but maybe it shouldn't be attached to the internet, for example, where it can take control over media and banking and voting systems? Can we think through the implications of our tech before the disasters happen for a change.

A weapon is a tool. A tool becomes a weapon. Fight! Fight! Tooth and nail - to the bloody end. "The end my friend." So lets keep fighting with each other because that's what we do in Christian America. AI won't teach us any different. It is just a tool to bludgeon another with or to extract meat from.

Doesn't take much computer at all to scream "Soros" and "you socialists!" That's what passes for "debate" these days, on the Luntz-2000.

why?

This article seems to report a veiled threat by IBM relating to U.S. Government 'soft professional controls'. Those controls, which are widely known in the Electrical Engineering and software professions, are used to control the development of artificial intelligence products and services. These controls generally take the form of export regulations attached to U.S. Government funding. In the article, the veiled threat comes in the selection of an Israeli student who apparently "argued against government subsidies for space exploration." Similar controls are applied in the electric power industry, to projects financed by the Dept. of Energy. In that case, technology for electric power generation equipment is sometimes patented in the U.S. by a foreign subsidiary of a U.S. company. Either the practice is unenforceable, or the U.S. Patent office looks the other way. The U.S. Patent office scrutinizes all incoming technology that relates to the DoD, DoE or NASA.

The real joy of human life is not about fiddlings with some clever gadget. The real joy and pleasure of life is in all the small interactions we have with people all day long. Almost all the interactive automation that I have seen is not for the benefit of people it is for the profits of large corporations. I want to talk to a grocer who knows me and cares about the fruits and vegetables he sells me. Not with some clever Amazon robot. I wish I could express this better...robots will not bring us a happy, caring and loving human society.

Apparently, my voicemail has already achieved intelligence; it passes the "two-ring" test.

When an AI system can understand a sentence like this one: Today I get to go to the store, and when I get there I will get what I need so I won't get sick, then I'll know we've really got something.

You know what this article tells me? We're still a ways away from achieving the Turing test.

I'll say this much: we certainly don't need AI writing up Wikipedia articles.

How about putting more resources to make sure our children get the education and training to develop critical thinking skills? AI is fine. RI (Real Intelligence) is needed a lot more now.

I don't really understand the goals off these AI programs. I'm not looking for debate partners or Jeopardy contestants. I use Siri on my iPhone and I'm satisfied if it can understand simple contextual questions. If I ask if I need an umbrella today, the AI can look at the weather and my schedule and determine if it's likely to rain while I'm out. The only actual product goal seems to be to allow companies to replace humans answering their phones with AI. And to do it in a way such that we wouldn't be able to tell we're talking to a machine.

Can we get this device installed in the Oval Office to replace the low tech device that currently occupies that space? It would be an upgrade.

The best short-term potential for AI is to enable a "bicycle of the mind" effect, to quote Steve Jobs. In other words, using AI to extend the capabilities of humans as they navigate tasks that would otherwise be time-consuming and mundane (such as making restaurant or hair salon reservations) or can be enhanced by being able to recall large amounts of information (such as participating in a debate). Most public discourse and media coverage of AI talks about threats to jobs and general intelligence, but the best short term upside potential we have is in assisting and empowering humans through narrow intelligence. We should be paying more attention to that.

Debate is not a good test for the Natural Language processing capability. In debates, one must take an opposite view from the opponent even if you don't believe in your own opposition. For example, if "Is Climate Change Real?" or "Is Earth Flat?" the person who takes the view "Climate change is False" and "Earth is Flat." would have painfully manipulated the facts to bolster the false conclusions! Only good use of this power is in political propaganda! Power of computers is in Machine Learning and giving a numerical weight to each argument with evidence. This is especially useful when the question does not have a definitive answer as is often the case in Medical Diagnosis.