Is AI above the law and going to destroy us because we don't understand it?

[rslade]

We've already wondered about the "help" artificial intelligence gives us.  And we've already wondered about legal issues, such as whether to trust forensic results that have been obtained through AI.

Now somebody is wondering if AI will destroy democracy.  (Actually, that's just a review of the paper, which can be found here.)  If that sounds a bit much, what he really seems to be afraid of is the concentration of AI power in the hands of some tech giants (a reasonable fear), and the seeming tacit acceptance that society and the law can't regulate AI because AI is above the law because we don't understand it.  (Although the "Google Spain" case that he cites seems to have been at least partly decided against that latter assertion.)

Personally, I think the paper is a bit wobbly, and hops from idea to idea without sticking to any one long enough to make a good case, but it does have some interesting ideas.

[j_M007]

Yes? No? Perhaps?

I noted with some wistfulness that the voice of HAL, Douglas Rain, has slipped the surly bonds of Earth. My "favouritest" AI there ever was.

"I'm afraid I can't do that, Dave."

https://www.bbc.com/news/entertainment-arts-46178930

[rslade]

> j_M007 (Contributor II) posted a new reply in Industry News on 11-12-2018 03:07

>  Douglas Rain, has slipped the surly bonds of Earth.

Appropos of nothing, I note that Douglas Rain was Canadian, as was John Magee,
the author of "High Flight."

====================== (quote inserted randomly by Pegasus Mailer)
rslade@vcn.bc.ca slade@victoria.tc.ca rslade@computercrime.org
It is a truth universally acknowledged, that a single man in
possession of a large fortune must be in want of a wife.
- Jane Austen (1775-1817) Pride and Prejudice (1813)
victoria.tc.ca/techrev/rms.htm http://twitter.com/rslade
http://blogs.securiteam.com/index.php/archives/author/p1/
https://is.gd/RotlWB

[j_M007]

As was is were,  James Doohan, William Shatner, and the Canadarm!

[DAlexander]

As is Justin Bieber

[Caute_cautim]

The real issue, is as usual human fear.   We do not have true Artificial Intelligence, we just dont have the computing capabilities or the resources to do this right now - AI is in fact Augmented Intelligence via Machine Learning and Cognitive Computing techniques.  What we have embarked on internally is identifying the number of different bias a human being introduces via our education, upbringing and intrinsic values.

If we can identify these, and then remove them - then the law firms would find it most usual indeed in terms of providing objective unbiased recommendations to assist a human judges or lawyers decisions.

Yes, we are closer to hacking the human brain or influencing an outcome - but what are the intrinsic benefits - one of the major ones is that people want to live to 100, and have their GP call them rather than the medical practitioner be sought to reduce costs and response time.

We have to weight up the whole picture, and not the hype that surrounds it.

Overcome the fear, seek out the information and make an informed choice.

Regards

Caute_cautim

[j_M007]

Meanwhile, in Gaithersburg!

source: SANS NewsBites, Vol. 20 Num. 090 

NIST Plans to Move to IBM’s Watson for Vulnerability Scoring

(November 2 & 12, 2018)
 

The National Institute of Standards and Technology reportedly plans to move to a vulnerability scoring method that uses IBM’s Watson artificial intelligence (AI) system by October 2019. The new system will take over the work now done by human analysts. In a pilot program earlier this year, NIST used Watson to examine hundreds of thousands of existing CVSS vulnerability scores; Watson then used that information to score new vulnerabilities, which were by and large in line with human analysts’ findings. Where Watson stumbled is when evaluating new and complex vulnerabilities; in those cases, human analysts will evaluate the vulnerabilities.
 

Editor's Note

[Pescatore]
I wonder how they know when Watson “stumbles” without an actual human person looking at every score?? That said, the Base and Temporal scores within CVSS lend themselves to algorithmic scoring, at least where everyone uses the same terminology – which isn’t the case in a lot of consumer products. The highest value part of the CVSS approach is the Environmental Metric group where each organization has to add a factor based on its IT environment and business impact potential. That factor is key in bubbling the most critical patches to the top of the list.

Read more in:
- www.nextgov.com: NIST Teams Up with IBM’s Watson to Rate How Dangerous Computer Bugs Are
- www.scmagazine.com: Report: NIST to use IBM’s Watson AI system to score vulnerabilities

[Caute_cautim]

@j_M007  Once again, some misinterpretation, it uses Machine Learning (ML); to look and analyse the traditional CVSS vulnerability rating and augments it by taking a different perspective - it takes a feed from the current point in time threat intelligence and then adds a weighting as to what is really happening rather a static view.   It works within the IBM X-Force Vulnerability Management System, augmenting the traditional vulnerability scanning techniques and takes a look at the real world.   It actually prioritises, which vulnerabilities need patching by the current state of the threats hitting organisations.  Some people call it weaponisation, by prioritising which ones need to be applied immediately etc.  In other words, attempting to be proactive or one step of the cyber game, which the majority of us play on a daily basis.

https://www.ibm.com/security/services/vulnerability-scanning

Then look at the Garner report, associated within the link as well.

Once again, Fear, Uncertainty and Destruction (FUD).....

Regards

Caute_cautim

[Caute_cautim]

Plus the fact, that is is a Machine, however it learns far quicker and faster then we can, and also it never forgets unlike human beings.