News, tips, advice, support for Windows, Office, PCs & more. Tech help. No bull. We're community supported by donations from our Plus Members, and proud of it
Home icon Home icon Home icon Email icon RSS icon
  • The promise of AIs to help people with disabilities – and a word of caution.

    Posted on OscarCP Comment on the AskWoody Lounge

    Home Forums Tech Accessibility The promise of AIs to help people with disabilities – and a word of caution.

    Viewing 0 reply threads
    • Author
      Posts
      • #2175927 Reply
        OscarCP
        AskWoody Plus

        The following article published today in the BBC World News’ site begins by explaining something that holds a promise of greater autonomy for people with bad eyesight or completely blind: an “intelligent” cane that contains motion sensors to determine its position and the wearer’s direction and speed of motion, with infra-red or acoustic emissions to sense obstacles in the direction the person is walking and capable of providing appropriate warnings and perhaps useful recommendations for a way around such obstacles.

        https://www.bbc.com/news/business-51632840

        The article continues discussing the advances in technology and algorithms that are making possible to have increasingly more “intelligent” devices, potentially, very useful to many in many daily situations.

        I do have a minor quibble of the author calling the many “neuron” that form the layers of self-learning machines “algorithms”, while the actual algorithm is the set of rules and principles implemented in the code that drives the whole AI system. The “neurons” are a very important and characteristic part, but only one part, of such a system’s driving algorithm.

        Over the last several decades there have been, after many years of trying various approaches, some serious advances in both software and in the theory and practice of learning systems, whose practical use is becoming possible thanks to the large steps forward taken in recent years, both in computer hardware, in software and the corresponding number-crunching capabilities they have, together, made available.

        Particularly impressive is the capacity to understand natural language, that makes it possible for a device to listen to requests and commands from humans and take correspondingly appropriate action, or for things such as “Google Translate” to work as well as it does now.

        But a word of caution is also in order here: while so-called “artificial intelligences” can be used now days to do much more than what were able their earlier versions twenty, or even ten years ago, they still are fairly limited and rather crude as “thinking machines” and tend to fail their assigned tasks more often that one would like to see them do that, and in many cases by making equally crude mistakes.

        Things of potential serious consequences to individuals, such as a misfiring AI executing a face-recognition algorithm to identify law-breakers (or regime opponents) in a crowd, are a real and increasing cause from concern. As can be having AI-based medical diagnoses used to determine actual patients’ courses of treatment.

        I have watched this field develop, with considerable interest, although it is not mine, since the early days of the perceptron of one layer, back in the late 60s, to the present adversarial learning “AI” consisting of a pair of them, where one creates problems that the other is supposed to solve, then the tables are turned, so now the other presents the problems and the first one has to solve them, and so on, together forming a very powerful “self-learning” algorithm.

        And, while following such developments, I have not only marveled at the positive achievements made, but been also often dismayed by the hyperbole with what many promoters of this field have been “selling the lion’s skin before hunting the lion”, as an old saying, in one of its various forms, goes.

        So people working in this field in its early days, back in the 70’s, were already predicting that machines capable of understanding and even translating natural language were just a few years from becoming reality (and I was using an IBM 360-50 with 600 MB of memory shared with several other processes besides my own, the one and only computer available to me and other students, to do the calculations I needed to get their results, before being ready to begin to write my doctoral thesis… so I was rather skeptical about that.) Well, they were right, but just only some four short decades off in that prediction. As was also the case with several other very impressive things prophesied, concerning what has become to be known as AI, particularly from the mid-eighties on.

        Windows 7 Professional, SP1, x64 Group W (ex B) & macOS + Linux (Mint)

        2 users thanked author for this post.
    Viewing 0 reply threads

    Please follow the -Lounge Rules- no personal attacks, no swearing, and politics/religion are relegated to the Rants forum.

    Reply To: The promise of AIs to help people with disabilities – and a word of caution.

    You can use BBCodes to format your content.
    Your account can't use Advanced BBCodes, they will be stripped before saving.