Either way, we’re all just strapped in for the ride.
1 Like
Trouble is, I’d guess the majority of people on the planet aren’t strapped in…
Yes, that is actually quite true considering that humanity as a race was at its strongest and healthiest during the hunter gatherer period 
Gone is the statement the camera never lies. 
5 Likes
Whilst I agree the sentiment it does depend a little on your definition of what is a camera. In days of yore cameras had no software and so could not be influenced by AI. As cameras were developed software was built in and steadily given more and more capability until we get to today’s cameras where much of the photograph is dependent on software algorithms. I’ll let others highlight where these algorithms change from being simple software to the more advanced software that can be labelled AI.
Well, yes - many many years ago; even when I were a lad, really.
Long before software, the camera could lie quite easily. At the simplest, you can make a small group of people look like a large crowd, simply by choosing your lens and camera position. There were many darkroom techniques to modify your pictures. Infra-red film dramatically changed your photos. As did filters. And so on.
2 Likes
Ai existed for a long time already:
I presume this is meant tongue in cheek - human lifespan is now more than double what the archaeological record suggests was the norm then. Humans are also much taller reflecting better nutrition…
AI is set to become a hugely disruptive general purpose technology. Hardly any industry will remain unaffected.
Some amazing stuff (outside of the chatbots and video creation examples) has already been achieved…
The average lifespan was shorter during the hunter gatherer period but that was primarily because it was more dangerous to live then, that’s right. On the other hand, people were both taller and physically stronger and less sick than when they transitioned to a farming society. The hunter-gatherer societies have been called by Marshall Sahlins the “original abundance society” because they usually do not need more than 4-6 hours per day and active individual to fulfill their needs for food, clothing and housing.
1 Like
Also the average lifespan was shorter because of the high infant mortality. People talk about an average lifespan of 40 or 45 years in the 1700s and 1800s - but go to any graveyard and look at the ages of the people who died. Either infants or 70 or 80-year-olds mostly. If you survived past young adult then the chances are good that you make it to old age.
3 Likes
As an occasion graveyard browser I’ve noticed that too.
I agree with Simon and others that what is now called AI did not suddenly come into existence recently. I was working on the applications of neural networks over 30 years ago which now fall within the umbrella of AI or machine learning. What has happened over the intervening years is the development of much more powerful computers and better algorithms which means that AI has become more available to the general public and consequently raised concerns among people who knew very little about the technology until recently.
2 Likes
Indeed, I notice you and I seem to have rather similar viewpoints. The other advancement is the far greater prevalence of big data, the WWW, and other public data sources with which to train systems.
Agree with your views in general.
What is perhaps more worrying is how these large data sets are compiled and managed.
Also what controls does the individual have over their data? Generally in order to gain access to a service one has to give (potentially) more personal data than is required to deliver the service being sought and in the process lose control as to what the service provider can do with it.
One area I am trying to get a better understanding of is organisations selling data to other organisations. They may say it is anonymised but it is simple to use analytic tools to quickly drill down to a point where individuals can be recognised especially if the person doing the drilling has access to more than one large data set.
Agh yes… this is one of my interest areas in my professional world, inherent bias in data sets which when used to train systems transfers that bias. Because there is no actual conscious intelligence in AI systems this issue can become problematic.
1 Like
Yes there are bias issues which need to be resolved.
There are also concerns around copyright and use of personal data. Some movement on this already as can be seen by large NY newspaper taking AI organisation to court regarding use of their copyrighted information.
Yes there is a difference between training and plagiarism. If the data is published on the public internet not behind walled gardens or pay walls then it is public domain… but it doesn’t mean you can plagiarise it, which is where the issues lie, but you can train with it… which is just like a person reading it to form an opinion.
With regard to bias on the WWW in terms of volume you have a disproportionate view of opinions… which leads more to sensationalism or extremist views (left, right and centrist) which tends to exaggerate negativity. Also key social demographic groups are over and under represented.