Four Predictions for AI

There has been much discussion on how artificial intelligence (AI) will shape our future. Questions like:
Will autonomous AI-enabled machines with sleek human form be a negative and unbridled force (Terminator style)? And will the development of AI improve the wellbeing of people, businesses and our community for the better? So let’s begin with four trends in which AI make grow between 2030 and 2050.

00:00 - Intro
00:20 - AI Discussion
00:43 - AI Prediction One
02:12 - AI Prediction 2
04:05 - Third AI Prediction
06:04 - Fourth AI Prediction
08:01 - Final Thoughs on AI
08:38 - Outro

Follow me:
Twitter @djware55

Licensed under Creative Commons: By Attribution 4.0 License

Licensed under Creative Commons: By Attribution 4.0 License

Werq by Kevin MacLeod

Industrial Cinematic by Kevin MacLeod

Music Used in this video
Licensed under Creative Commons: By Attribution 3.0 License

#AI #Predictions #2030

[CyberGizmo]

I escaped the bot...and will be back in the video - good thing I watched the Houdini Movie. Oh, and yeah this video was made entirely on Linux.

[DanielSMatthews]

The knowledge required to live a long and healthy life full of happiness and meaning has been around for several thousand years, yet we see conditions and life expectancy in the USA currently falling as technology rises therefore we conclude that only the introduction of "slap drones" to make people do the right thing by themselves and others will improve the situation significantly.

[abobader]

I was hoping to hear the voice of SAL :) Very nice DJ :)

[tonywise]

I'm too old to live to see most of this (75), but my worry is system security. Unless we can trust that systems cannot be compromised, AI may prove more of a danger than a benefit to the human race.

[RafaCoringaProducoes]

DJ antecipating the next q&a, what do u think about copilot, libreoffice should implant something like it? By the way i see copilot integrating with win12 ecosystem, projects like chatgenie could aim this way?

[dezmondwhitney]

Not up to speed on AI. However my gut feeling is a Real Concern about the enabling of those without any scruples to Exploit naive and trusting, non- IT, people and to get away with it using plausible deniability of some sort.

[JonathanSteadman]

Cant wait till robots program themselves.

[leibaleibovich]

Thank you for the video!
Very interesting takes on the future of AI. I would like to offer my controversial opinion, if I may.

I am tangentially related to medical field in my current job. A great deal of medical diagnostics depends on "visual" image "processing", so to speak. Think of a radiologist assessing an X-ray image. Think of a pathologist, who under microscope examines a section of malignant tumor to understand if cancer has been completely excised from the body or a second surgery is needed. From AI standpoint, the problem could look similar to image recognition / classification / pattern detection. It should not matter, if we are trying to classify cats on home photos or histo-types of invasive cancers on H&E microscopy slides.
If memory serves, there have already been attempts at developing an AI software for diagnostics of lung cancer using digital images of microscopy slides.
However, in my opinion, the problem lies in a different plane. All medical doctors (including pathologists) are covered with different types of "malpractice insurance". Every pathology report has a name of pathologist, date, full histological description, diagnosis etc. All microscopy slides, used for diagnostics, are stored in hospital archives. They can be retrieved at any time and sent for a second / third opinion in case of litigation. In case of mis-diagnosis, the pathologist will be held liable, hence, the need for the insurance.

In case of a diagnosis (or a wrong diagnosis, should I say?) made by AI, who will be held liable? Software developers / vendors, the hospital, the medical practitioner, who uses the AI? What will insurance companies say? Questions of liability and malpractice are central to medical profession. In case of AI, there are no clear-cut answers to these questions. At least, I am not aware of such answers. I would speculate that legal conundrums might pose bigger obstacle to implementation of AI in healthcare than technical difficulties that still have to be surmounted.

And I have not even touched genomics and all problems that might be associated with use of DNA information.