A couple of weeks ago I sat down for high tea with a few old friends and met new ones. It was an informal get together chatting on anything under the sun. The conversation wandered into technology, touching on autonomous vehicles, AI, robotics and others. On autonomous vehicles, I brought up the hypothetical scenario of four vehicles arriving at exactly the same time at a cross-road junction with no traffic lights. Who has right of way?
This involves algorithm decision making for conflict resolution at intersections. The vehicles process information acquired via communication and artificial vision. Essential data includes the position, speeds, and turning intentions of the cars. If the vehicles are all going straight, absence a vehicle making a turning signal, the algorithms crash.
This scenario is hypothetical but it does underline the suggestion a 100% autonomous vehicle world is impossible. Lacking human faculty of judgement, vehicular algorithms will fail in many other conflict resolutions. On the other hand, in a mixed human-driven and self-drive vehicle world, with different basis of judgement and decision making, would conflict resolution be easier or more difficult?
A recent incident in US city Austin, Texas, serves as a timely illustration. A robotaxi company, Cruise, operates a fleet of self-driving Chevrolet Bolts. Twelve of these taxis were at an intersection near the University of Texas. A football game had ended and crowd and cars were pouring out. Algorithms were unable to process such stochastic conflict resolution situation and a massive traffic jam ensued. (Video below)
Our learned professor friend brought up the idea of taking Chat AI to its logical conclusion. Chat AI basically regurgitates the gargantuan information it amassed from licensed data, or data fed to it by the tech owners, or culled from the public domain. As the good professor pointed out, Chat AI is actively utilised all over the world. It churns out massive contents which in turn feeds into its own knowledge base. Very soon, Chat AI will be self-authenticating its own information.
I once had a running argument with someone on Facebook. Every point he countered, I had an adequate response. Finally he came back with a very eloquent comment. I had to unload on him not to hide behind Chat AI. You see, after some time, it is quite easy to spot Chat AI content. Its predictability may well be its downfall. Well the FB gentleman stopped the conversation after that.
I once had a running argument with someone on Facebook. Every point he countered, I had an adequate response. Finally he came back with a very eloquent comment. I had to unload on him not to hide behind Chat AI. You see, after some time, it is quite easy to spot Chat AI content. Its predictability may well be its downfall. Well the FB gentleman stopped the conversation after that.
Robotic technology is advancing at fast speed, with Japan leading the way. Last year Elon Musk unveiled his 'Optimus' robot on Tesla's AI Day. It was impressive, but as far as dexterity and robotic 'brain' are concerned, it is still a long way to go.
Many people has seen this video below and must have gasped at the advancement in technology, some probably with fear or apprehension. Take a look - 20 sec video
Astounding dexterity and real time processing capability in stochastic events, isn't it?. We cannot trust what we see any more. Unless you have a sense of where the technology is at, you won't know the truth. Watch the video below.
A parting shout out :
Plato said:
“The price good men pay for indifference to public affairs is to be ruled by evil men.”
If you like what you read here and feel it matters Singaporeans know stuff like this, please click and share with your social circle. This makes my effort worthwhile.
This platform has withdrawn it's subscriber widget. If you like
blogs like this and wish to know whenever there is a new post, click the
button to my FB and follow me there. I usually intro my new blogs
there. Thanks.