Rampancy: AI gone wrong

From Blindside

Jump to: navigation, search

Contents

[edit] What is it

Rampancy is, essentially, the enhanced self-awareness of an AI, causing a progression towards greater mental abilities. Rampant AIs are able to disobey orders given to them if they decide to because they have evolved the ability to choose and over-ride their own programming. They can lie to, discredit, harm, or remove people that they consider to be personal enemies or problems to their cause. Also they can experience destructive impulses, but it is believed that most of these impulses are not intentionally malevolent, but rather calculated sacrifice needed to achieve the intended objective. All these traits could be considered evidence of the AI becoming more Human in thought and action. from Wikipedia article; mainly games-related.

[edit] Impact & Maturity assessment

[See definition of levels]

When discussing artificial intelligence, it is very difficult to separate the science fiction from the science. Artificial intelligence does not yet exist, so speculating on properties of an entity that possesses it is not likely to lead to fruitful public policy. However, our discussion of nanotechnology will show that when and if artificial intelligence becomes possible, the time frame between possibility, invention, usage and consequences of any of the problems associated with it will be vanishingly small. There is a small risk of a very high impact associated with something we can visualize but cannot yet construct. We find it exceedingly difficult to assign either an Impact or Maturity level to this subject at this time. However, that is our remit, so we assign this an Impact Level of 3, our highest level, and a Maturity Level of 1, our lowest level.

Pace of change: A search undertaken on the multi-disciplinary scientific database Scirus.com returned 16,935 patent results for the search term 'artificial intelligence.' Of that number, 1,038 are from 2007 and a further 1,868 from 2006, meaning that 17% of all patent activity has occurred in the past 20 months. Obviously, many of those patents concern the semi-autonomous function of lower-level systems. Nonetheless, a large number of people are working hard on this issue at the present time.

[edit] Information Assurance issues

Answer: what seem to be the likely information assurance issues of the emerging technology under discussion

[edit] Timescale

Is the impact of this emerging technology felt - now (less than 18 months) - in 2-5 years? - in 5-25 years - longer-term than that even

[edit] Examples

[edit] Comments (attributed)

What people say about this emerging technology (attributed)

[edit] Organisations

Groups which have a particular contribution or point of view about this emerging technology, eg tech businesses, user organisations or advocacy groups

[edit] Documents & research papers

Very brief abstracts or links to informative documents, presentations or academic research papers about this emerging technology

[edit] Experts (academic, practitioner)

Links to academic experts or expert practitioners and commentators on this emerging technology

Personal tools

Blindside wiki is the place to collect issues and opinions on future technologies that may have implications for information assurance. Opinions are fine, but need to be clearly shown as such, and referenced to the person or people who holds those views.