Ready or Not, We’re Entering an AI World

The tech landscape as we know it is about to be obliterated.

No, I’m not talking about the impact of a Trump presidency, but something bigger—much bigger.

The impact of Artificial Intelligence (AI) and related technologies, including machine learning, deep learning and neural networks, is now starting to be felt almost everywhere we look. (See my previous column “Learning About Deep Learning” for more.) From everyday devices like PCs and smartphones, to media companies like Facebook, to smart home security services, to connected cars and even to medical diagnostics, the influence of AI is growing rapidly.

Not only is it changing the products and services we use, it’s dramatically reshaping the technical infrastructure behind cloud-based services ranging from web searching to communications bots to personalized advertising (unfortunately!) and much more.

Driving improvements in AI performance is also becoming a key motivating factor for product developments and strategic partnerships. In fact, it seems like virtually all the biggest news-related announcements from tech vendors now have some kind of AI-related angle.

On the semiconductor side, for example, nVidia, AMD, Intel, Qualcomm and many others are focused on creating chips and software that can drive massive improvements in deep learning in large data centers. At this week’s SC16 supercomputing conference, for example, nVidia talked about its efforts on the Cancer Moonshot, a project it’s working on with the National Cancer Institute and the US Department of Energy to deliver a decade of advances in cancer research in just five years thanks to GPU-driven AI applications.

AMD unveiled a new partnership with Google for using its latest GPUs in Google’s Compute Engine and Cloud Machine Learning services, as well as enhancements to its open-source AI-focused Radeon Open Compute Platform (ROCm).

In the world of media, Facebook just announced that it’s planning to leverage AI to help combat its problem with fake news stories. In the devices world, Apple has talked about its AI efforts to help make Siri smarter about your needs and interests, while still maintaining privacy. Google, meanwhile, is leveraging AI to build up a whole range of services that are both more contextually aware of the environments you might find yourself in, as well as personally aware of the specific ways you like to use them.[pullquote]The question is, are we ready for the onslaught of AI? It’s clear that tech-related companies certainly are. But for individuals, the answers are a bit more opaque.[/pullquote]

In short, we’re being surrounded by the first significant fruits from the long-growing but previously near-barren AI tree. For many, these fruits may not offer much taste yet, but it’s clear from early nibbles, that we’re due for an explosion of flavor.

The question then is, are we ready for the onslaught of AI? Based on numerous signs, it’s clear that tech-related companies certainly are. But for individuals, the answers are a bit more opaque. Sure, it’s easy to get excited about the possibilities that AI and related technologies offer, but there can be downsides as well, particularly regarding privacy.

The real trick behind many AI-based products and services is to recognize patterns and then react to those patterns based on previous knowledge about an individual’s preferences. In a positive case, a digital assistant might be able to use that knowledge to help you make better decisions.

As many have started to recognize, however, technologies with good intentions can often be used in unexpected and negative ways. The same set of data about your habits and preferences could be leveraged by criminals to figure out how and when to digitally burglarize you, for example.

Unfortunately, understanding not just the intentions but the decidedly human biases that creep into (or even form the foundation of) the algorithms that drive AI-based products can be very challenging. Nevertheless, as the AI era dawns around us, it’s best to be prepared for a wide range of potential outcomes, with the knowledge that there are bound to be few unpleasant bumps along the way.

Published by

Bob O'Donnell

Bob O’Donnell is the president and chief analyst of TECHnalysis Research, LLC a technology consulting and market research firm that provides strategic consulting and market research services to the technology industry and professional financial community. You can follow him on Twitter @bobodtech.

13 thoughts on “Ready or Not, We’re Entering an AI World”

  1. “The real trick behind many AI-based products and services is to recognize patterns and then react to those patterns based on previous knowledge about an individual’s preferences.”

    This sounds like it’s all backward looking and big-data based which sounds like converge to the mean, or worse, devolve to the mean. That’s fine as long as we realize that an AI decision engine is not something that we should rely on for leaps of the imagination. We need to be smart about where AI makes sense and where it doesn’t.

    I’ve always believed that the information revolution is leading to a bifurcation of most of society into the people who ‘push a button’ to make a living and the people who ‘make the button work’ for their livelihood. [There will be that very small but very wealthy segment that owns all the buttons.] Which side of the button you work in will determine your standard of living.

    AI is likely to exacerbate this bifurcation because that means the skills, especially cognitive skills, required of people who push the buttons will get lower and lower. Not only that, the number of buttons to be pushed by humans will also decrease as machines get smarter and smarter. Both of these trends all put negative pressure on the incomes of the less skilled segment.

    These are tricky long term issues that society needs to think about. The number of people who are angry, poorly paid, unemployed and unemployable, and undereducated (relative to the needs of the economy) people will get a lot bigger if we don’t figure this out.

    1. “…people who ‘push a button’ to make a living and the people who ‘make the button work’ for their livelihood”

      Yes, and at the top will be a mechanical piano that plays pre-programmed tunes. And we will elect the DJ for it. Some questions that remain to be asked: who composes a new music for the piano? who develops a good taste for the music for the general population? who makes sure that DJ does not sleep?

    2. I’m not sure tech evolution has much to do with social justice. The haves vs have-nots issue has been around for a long time, its parameters more or less understood (capital vs work, social safety net, upward mobility, redistributive taxation…). Whether it’s about making AIs, making cars, making crops, controlling the hunting grounds… same-ol’, same-ol’. Not even sure the issue is getting sharper.

      1. I disagree. Technology has a lot to do with a social justice and humanity. For example, at a new hire orientation Apple portrays itself as standing on a crossroad between humanity and a technology.

        There is a good TV series on NetFlix “Black Mirror” (England) which tells the story where technology can lead us if we are not careful.

        1. Agreed, technology has a lot to do with that. But
          1- It’s arrogant and myopic to think today’s tech has more impact on it than yesterday’s or tomorrow’s. Why AI and not black powder, medicine, books, fiat money, telecoms, computers…
          2- things besides technology have a lot to do with that too: politics, education, ethics, social mores…

          1. Technology’s effect on social justice is always about society’s transition to accommodate or adjust to changing technology.
            We tend to think that society has been able to weather these changes quite successfully, just look at where we are now, we’ve gone a long way since the industrial revolution. What we seem to overlook is, that ‘weathering these changes’ typically means just waiting for the generation of displaced, unsuitably skilled people to die out. When they die, then the problem ‘disappears’ and social justice is restored.

    3. Since the information revolution is already happening, you can already test your hypotheses.

      My observation is that although there has been a widening of income difference, it doesn’t map to those pushing buttons/making buttons. That is to say, people making buttons aren’t necessarily rich, and people pushing buttons aren’t necessarily poor. In fact, button pushers like sales reps, construction workers, doctors, managers make good money. On the other hand, button makers like the poor souls developing websites or developing apps for App Stores don’t make too much unless they have something special to offer.

      I think it just boils down to demand and supply. Jobs where supply outstrips demand don’t pay well.

      There are many jobs where demand hugely outstrips supply, and for which AI is still very far from being able to fulfil. Caring for small children and the elderly are areas where, in Japan, demand is huge but supply is small, partially because the huge labour intensity required doesn’t match healthcare budgets. Although completely replacing these jobs with AI is still decades away, any innovations to make this service cheaper to provide, could dramatically improve the economics and make it a feasible business even without government subsidies. This would hugely expand the market.

      I look forward to an AI world, because I see our lives being enriched for both the rich and the poor.

      Current day economics of tech suggest that the benefits of AI will not be restricted to the rich (which would be terrible), and that AI will actually help to level the playing field. It will bring affordable customised education to those that cannot afford private tutors, etc.

      1. I hope you are right. But you are talking about the present, and near term future. I am talking about the future that the Robotics/AI sector, based on all the sounds emanating from them, seems to be working towards. They talk about robots and intelligent machines to take over, perhaps not completely but significantly, a lot of service industry jobs in care-giving, healthcare, restaurant service, transportation services, defense, retail, etc. Put that on top of the manufacturing jobs that will continue to disappear, where will all these people go to find sufficiently gainful employment? I fear that problem will be ‘solved’ the same way history has always dealt with masses of displaced workers: wait for them to die.

Leave a Reply to Naofumi Cancel reply

Your email address will not be published. Required fields are marked *