Everyone talks of machine learning, AI and automation these days. Automation could be disruptive and could have far reaching social implications. Should there be limits to automation? and who decides the limits to automation- scientists, corporates or states and its people in general?
Good question. My first thought is that any limits that may be set by any party will always be overcome by perceived military requirements. Hopefully this is too pessimistic...
JD
James, are you suggesting that 'state' is the agency that should define the limits? Because it is the 'state' which perceives the threat.
Veerendra -- my thought is that it may well be a state that, perceiving a threat, authorizes its military to overstep set limits. Or the military might do it autonomously.. . Or a "terrorist" organisation. It all depends on how high the stakes are?
James
On the conceptual level I would say scientists "produce knowledge" so it is not their primary job to define limits. Stll, they might help identifying risks. Decisions will have to be taken on the policy/governance level which will include relevant societal actors (industry, "people"...). What do you think?
I agree Ms. Stefan Lasser that scientists may help arrive at the decision, but the decision needs to be taken by the agency of 'State'.
Why not at the global level e.g. UN? Otherwise decisions will be taken competitively in the commercial/military interests of individual states -- to global detriment?
Consider e.g. the nuclear test-ban treaty which seems to have worked pretty well so far?
James
The following, quoted from a public domain BBC News web page of March 21 2017, is relevant to this discussion I think:
"Information Warrior 17 will involve Artificial Intelligence (AI) and test the protection of warships and submarines against cyber attacks.
"The training will be held during Joint Warrior, a major Nato exercise taking place in and around Scotland from 26 March to 6 April.
"This exercise, which is held twice a year, involves thousands of army, navy and air force personnel.
"The Royal Navy said Information Warrior 17 would "set the foundations" for cyber warfare in the future "because, as our enemies become more technologically advanced, so must we to combat these threats".
The exercise will involve:
■Using AI technology to develop a "ship's mind" at the centre of Royal Navy warships. The Navy said this would allow fast, complex decisions to made made automatically making warships and submarines "safer and more effective in fast-moving, war-fighting situations" (my italics)
...
"
So, the limits to automation will always be breached depending upon the threat perception and hence there will no limit at all other than the technological one. Paranoia will occupy the driver's seat.
I am sceptical that automation can be stopped or controlled. During the industrial revolutions in Europe thousands of acts of sabotage were carried out by Luddites and saboteurs (actually throwing sabots, a kind of wooden clog) into the machinery. Various laws and edicts also tried to prevent technological advancement to no avail.
We need to learn faster how to adapt to this evolution. The arrow of time moves inexorably in one direction, however much we yearn for the 'good old days' they ain't coming back.
Yes there should be limits. What is against the common good should be forbidden.
Automated collection of information on those using the web by corporation and government should be limited and regulated. We should have some privacy when we used the web. These are only a few examples.
Now, who should decide about these limits? It should be a societal collective decisions. When we have concerned about a particular new automation, there should be a public protection agency collecting all these type of concerns and that would analyse them and do further inquiries and then proposed new regulations to the legislators when this is required, i.e. when some of these automated practices is detrimental to the public good. This agency should be independent from politicians and corporate money. Ideally the governance of this agency should be elected and its funding should be protected.
As we know automation has direct reference to equipment , manufacturing ,& for researcher & user where he may used is own convenient facility or other process.It is in this line that it is not possible to offer the limit to any areas directly as it depends on the nature of process for the user of automation & the purpose in which his activities are carried out .
This is my personal opinion
- Badges
- Science topic
- Similar topics
- Social Psychology
- Attitude