Waves of AI change
South Curl Curl, Sydney, July 2017

Waves of AI change

I think we all recognise that adapting to AI use within our various organisations is going to be quite a challenge. To date I’ve seen less awareness of the fact that there will be waves of change, with each wave more impactful than the last. Here’s a broad outline of how I think things may evolve in the workplace:

First wave: AI as support tools

Initially, AI may be seen as a set of support tools, not conceptually all that different from spreadsheets or other productivity software. These AI-powered tools will be designed to enhance human productivity by automating routine tasks, providing insights, and assisting with decision-making. Employees will still maintain control over their work, with AI serving as a powerful aid to boost efficiency and effectiveness. This is relatively “safe” AI, the kind most often referenced by AI proponents.

The good: More productivity! Woo-hoo!

The bad: You snooze, you lose. If you, or the organisation you work for, can’t adapt fast you risk losing relevance in an eye-blink.

The ugly: Even in the first wave there will be mass layoffs in some job categories (eg. copywriters, content creators, customer service agents). Even in an ideal scenario where as many (or more) new jobs are created as old ones are lost this will take time, with plenty of angst during the transition.

Second wave: AI as colleagues

As AI becomes more sophisticated, it may take on roles more akin to workplace colleagues. AI entities will work alongside humans, performing specific jobs with minimal supervision. If you aim to put a positive spin on things you describe this wave as freeing humans to concentrate on creative, empathetic, and strategic roles, while AI handles data-driven, repetitive, and computational tasks. A less positive spin might note the increasing opportunities for a decreasing head count.

The good: A lot less of us will be doing boring, repetitive jobs.

The bad: A lot less of us may be doing any jobs at all.

The ugly: Paranoia much?

Third wave: AI as supervisors

With their inherent capacity to know and keep track of more stuff than any human can, AI may gradually assume the role of workplace supervisors. AI may simply become best-placed to direct human employees to perform tasks that optimise overall productivity and profitability. These supervisory roles may span all the way from determining production schedules and shipping logistics to high-level strategic decision-making. Additionally (and perhaps creepily), AI may be tasked with monitoring and assessing human performance, providing feedback and recommendations for improvement.

The good: Getting the basics right, consistently over time, is one of the keys to success for any organisation. AI supervision will (eventually) reliably deliver ideal execution, on demand.

The bad: It’s hard to imagine people accepting subservience to AI - except we already have. Algorithms rule our lives, just not overtly, yet.

The ugly: Because people will probably object to AI bosses it’s likely that reality will be obfuscated. They may not be above us in the org chart, but they’ll still be in charge.

Fourth wave: AI as ineffable oracles

As AI capabilities surpass human understanding, some AI systems may come to be treated as a form of "local god" within organisations. These advanced AI entities will operate in ways that are increasingly beyond the comprehension and critique of human employees. In such cases, the AI's decisions and actions may be accepted as mysterious yet infallible. This shift in perception could lead to a new dynamic in the workplace, where humans defer to AI for guidance and direction, trusting in its superior situational awareness and decision-making capabilities.

The good: For AI to gain this level of trust from us it will have to give us good reason to trust it - that is, it will actually need to consistently provide us with good guidance.

The bad: AI is probably going to get really good at subtly manipulating our perceptions (noting that this doesn’t require ill intent - it may be manipulating us towards “happiness”).

The ugly: Consider the story of the Land of the Lotus Eaters (the version from Homer or the version from Star Trek - either will do.)

Each of these waves of change will profoundly challenge our basic assumptions about how organisations should function, and about our own roles and relevance at work. I expect we’ll find ways to adapt over time but how fast are these waves going to hit? And how certain is all of this anyway? I doubt anyone can say for sure but:

  • Wave one is a given as it's already under way, albeit with products that are still flaky at times.
  • Wave two probably needs capable and reliable AI agents (i.e., AI that take a broad goal and turn it into sub-tasks they can implement under their own steam). Good chance this will be real in a commercially relevant way by later this year.
  • Wave three also requires agentic AI but in a limited way it's already here. In an enterprise the "copilots" that are beginning to help us code, and write, and organise our meetings etc. aren't working solely for us as individuals. When they remind us that we have a backlog of incomplete tasks, or helpfully point out to us that we tend to get short in meetings with Karen from accounting, we are not necessarily the only ones benefiting from those insights.
  • Wave four is tougher to estimate. In its most complete form it may depend on whether we can get to AGI/ASI (Artificial General/Super Intelligence) just by building bigger models or if we still need some additional fundamental breakthroughs. Might be in two years, might be in twenty. Perhaps we should hope for twenty.

However it all pans out in reality, if you’re an executive with change management in your area of responsibility I really hope you’ve had your Weetbix.


Note: Posts with images get more views, apparently. I’ve used AI-generated images in the past, which was fun for a while, but one of the things AI is doing is creating a world in which reality and authenticity are becoming luxury goods. Thus I now choose to use my own photos. I’m not claiming they’re relevant or good photos, but they are my photos, taken thoughtfully and with care by an actual person.

"The future is already here — it's just not very evenly?distributed." William Gibson

回复
Victoria Young

Managing Director, Privacy & Cyber Security at Scyne Advisory

10 个月

Interesting… If we’re destined for a world where human labour is increasingly devalued, what does the economy look like? Thoughts on universal basic income? Or do we all need to be equity owners in the AI systems generating value? I think, from a policy perspective, it’s going to be really hard for nations that have traditionally had strong social welfare systems and tax-based wealth redistribution to manage the kind of transition you’ve described. Seems like there could be significant risk of wealth concentration and competition challenges.

要查看或添加评论,请登录

Barry Thomas的更多文章

  • Not Everyone's Going to Get Good at AI

    Not Everyone's Going to Get Good at AI

    I've been watching how frontline staff engage with AI at some client sites, and I'm seeing something fascinating that…

    1 条评论
  • AI's Real Revolution: Raising the Floor, Not the Ceiling

    AI's Real Revolution: Raising the Floor, Not the Ceiling

    In my first programming job, everyone had stuffed toys on their monitors. The practice wasn't childish whimsy - it was…

    4 条评论
  • Dethrone the Phone: The Coming Shift from Smart Devices to Persistent AI

    Dethrone the Phone: The Coming Shift from Smart Devices to Persistent AI

    Watch any tech preview about AI and phones, and you'll see an oddly limited vision: increasingly sophisticated AI…

    6 条评论
  • When Objects Wake Up

    When Objects Wake Up

    It's a strange world, but it's about to get a lot stranger. While the public gawks at ChatGPT and its cloud-dwelling…

  • AI: Your Plastic Pal Who’s Fun to Be With

    AI: Your Plastic Pal Who’s Fun to Be With

    Here's something nobody seems to talk about enough: having a genuinely intelligent conversation partner available 24/7…

    2 条评论
  • To survive the rise of AI we need to re-invent ourselves

    To survive the rise of AI we need to re-invent ourselves

    AI is transforming our world—quickly, profoundly, and not always for the better. This transformation demands that we…

    3 条评论
  • 50% of people are below average

    50% of people are below average

    There's a cognitive trap I see many business leaders fall into when discussing AI. They focus intensely on whether AI…

  • Your website is an irrelevant waste of time and money

    Your website is an irrelevant waste of time and money

    Ok, that's a clickbaity title but let's be blunt: when was the last time someone actually wanted to browse your…

    5 条评论
  • Are You Wondering What AI Is Good For, For You, Today?

    Are You Wondering What AI Is Good For, For You, Today?

    I often find myself in conversations with people who assert that AI just isn’t all that practically useful yet. They…

    1 条评论
  • The Velocity Imperative

    The Velocity Imperative

    As artificial intelligence begins to transform the business landscape, velocity of action - that is, the ability of an…

    1 条评论

社区洞察

其他会员也浏览了