Ok, so AI will not give you more time (see part I). And AI can be this huge threat to mankind etc. – according to an increasing collection of experts. Sounds serious, but it’s still kind of distant, isn’t it? So let’s bring it closer to home: Is AI a real threat to your job – not 10 years from now, but this year? The answer is yes, but only if you’re lazy. Are you?
First – this is not another deep analysis of your professional future (or mine), just some practical reflections. Like the fact that all these incredible digital models, bots and services need resources – they run on something, come from somewhere. They ‘are’ software, live in datacenters, live off of data, drink lots of energy, need constant development, maintenance and tuning, they need front ends, interfaces, you name it. So if you have just about any computer science or operations background and are willing to adjust your professional trajectory according to market requirements, you’re all set. If you’re good at it, you’ll always be in demand.
For example: The new wonderbots will undoubtedly do a lot of programming, ah, well, sorry, software development, but someone will have to ‘feed them’ and manage them – set targets, run tests, check results, be responsible. That may be you and it will likely be a lot more exciting than churning out code. Also – and regardless of type and tasks – the bots need tons of data – collected, curated, checked, prepared by professionals that understand the process, the goals, the LLMs and not the least, the sources. A new generation of ‘data experts’ or ‘data scientists’.
Less obvious, they need controllers. Easy to forget, but honestly, the machines don’t know what they’re doing. Don’t understand. But you do. Off the bat it may seem like a boring job, but it is the opposite. Think about it, it’s research, tracking down patterns, understanding, connecting dots etc., constantly – presumably with an efficient feedback loop to developers, data scientists and more.
Also less obvious – and at the opposite end of the process, the user end if you like – is what I used to call ‘input engineering’, recently rechristened ‘prompt engineering’ (recently discussed in AI: And Now You’re a ‘Prompt Engineer’). While the general chatbots are (or should be) capable of making sense of human input, the specialized ones (actually most of them) will need help to understand exactly what’s expected from them. Specialists that know how to ‘formulate’ the input (like, a question) to get the desired (or best possible) output.
If you think that’s a temporary requirement because the bots are getting better all the time, here’s a different take: Development (aka evolution) is taking ‘AI’ to new levels and in new directions all the time. Including putting the bots into (physically) small computers to make them accessible/useful in new settings (think ‘bot in a weather balloon’ or ‘bot in a security scanner’). The UI is going to have low priority in these settings for a long time. If you’re a good prompt engineer with deep knowledge about the actual field or segment, your professional future is again safe.
The list goes on and doesn’t have to be technical in nature. Recycling is a great example. Discarded batteries and solar panels (among many other things) are piling up at an increasing rate, millions of them, containing lots of toxic and rare/valuable materials. Our smart machines will be invaluable in the growing recycling industry, but I guarantee you there are plenty of safe professional opportunities forever – meaningful, challenging, advanced. (Check out this interesting WIRED Magazine article about the subject.)
Finally, there is a lot of creative work to be done in developing new skills and new markets for the new machines. Pundits may disagree, but the machines are not going to do that themselves – not now, probably never.
The point is: As we put our new tools and helpers to work, new opportunities are created at about the same speed, maybe faster. Which means that unemployment is a choice, not a mandate. And definitely not a threat.