Petition to pause AI experiments

I am referring to a news item on NDTV where Elon Musk and many others have asked for a six-month pause in all Major AI developments such as the next version of ChatGPT by Open.ai.

https://www.ndtv.com/world-news/elon-musk-experts-call-for-pause-in-giant-ai-experiments-3903561

The key message from their letter which I found profound is, we get only one chance to build great intelligent machines and hence we need to be careful about putting sufficient safeguards built into the machine.??They basically implied that, once you have built the first true AI machine, then it will take over and we may not get a chance to put the genie back in the bottle.?Scary, especially coming from Elon Musk who has gone more advanced in his Tesla autonomous cars than anyone else.

AI&Ethics is my favorite topic in the AIML area and this letter is a shot in the arm for all of us who have been worried about this issue of AGI machines being built by various private and governments while the policy teams are very slow in drafting and adapting ?Ethical AI laws.?The EU has taken an early lead but being a multi-country organization, it takes q while to get from the draft stage to the fully implementable law stage.?Many countries are in the initial stages of expert committee formation, and initial drafting stages.?When ChatGPT version 3 was released to the world in November 2022, it never went thru any external ethical reviews.?In less than six months, the next release is out and a multi-Billion dollar startup ecosystem is formed surrounding generative AI technologies.

However, I want System Integrators and enterprises NOT to Pause their AI programs.??Many of the AIML use cases in Factories, warehouses, supply chains, and customer management are in the ANI space where the learning machines are asked to learn a particular narrow domain.?However complex the domain is, the ANI machines are self-learned expert systems e.g. Visual quality control inspection agent at end of the assembly line. ?The QC machine knows what to look for in the final product to identify faults and works hand-in-hand with various other devices such as cameras, measurement devices, IOT machines, etc.??The QC Robot holds and rotates the product at various angles, taking measurements, etc. will be part of the QC machine. The machine learning of recognizing minute flaws that are built over decades of human experience is learned by the machines by actual data collected during the human QC era.?However sophisticated these systems are, the machine knows only one domain.?One product, one QC, and can only output product Pass or Product Fail. ?These are ANI Artificial Narrow Intelligence machines.?

No need for enterprises to stop any of these ANI Business use cases.?Please press the pedal as the advances in AGI will also help in making ANI machines more sophisticated.

The worries raised by Elon Musk and others are for attempts at AGI Artificial General Intelligence such as ChatGPT etc.?Even here many critics feel Elon has raised the red flag too soon and there is no cause for worry now.?I am sure lots of self-regulation and checks & balances are being done by open.ai and others without any government intervention. ?Governments will soon come out with safeguards.

Wishing all the Enterprise AIML projects great success and urge System Integrators to reskill their staff in quickly harnessing these technologies to their customer’s benefit.

#aiml #aiethics AiThoughts.Org