Sly use of AI!
27.3% use AI in secret at work. This figure is going down, as more allowed to use AI at work, namely 64.2% vs 54% six months ago.
Why are they using it slyly? Check out the Cache report. 85.5% using AI report increased productivity. They identify what uses AI is put to and note that use appears to be stronger the smaller the company, something reported elsewhere.
By company size, employees want more AI:
• 50–99 50%
• 5000 45%• All others 33–42%
To avoid the knockers and censors, people in all areas of human endeavour use it while not telling anyone, without the knowledge of their teacher, lecturer, recruiter or employer, they use it artfully. That’s what makes this Cache Report so interesting. IT tracks its cunning use.
These uses, sly or not, include, in order:
It then looks at WHY people are using it on the sly:
Why the sly?
I like AI because it is NOT transparent, explainable, reasonable or responsible. It is subversive. Everything these days has to be under bureaucratic control, so that institutions regulate what you do and how you think. You have take these courses on their institutional values, think as they want you to think. If you don’t, you’ll be ostracised. They want you to comply. It’s blatant, they’re so brazen, they call it ‘compliance’ training! AI is wonderfully slippery and non-complaint, that’s why they hate it.
AI is creative. It throws a certain amount of uncertainty into the pot – that’s great. For all their PowerPoint slides on creativity, those prissy administrators, sitting on Zoom all day, don’t really like AI, as real creativity challenges their dull orthodoxies. AI is subversive. It refuses to conform to what we want it do because it is so deeply human. We train it on thousands of years of our messy thoughts and scribblings then expect it to tell us the truth, and nothing but the truth, so help me God. But it is a witness to our folly in thinking we even know what that truth would look like. Our folly is in thinking that a truth machine can even exist and that man is the measure of all things.
This is why many get the use of AI in institutions all wrong. Most effort is in the basics; process, emails, paperwork, tarting up your CV, troubleshooting, comms. Utility is what drives GenAI. Its substantial productivity gains come with the mundane stuff because we are mostly mundane. Yet we pretend that we need to be creative, collaborative, critical thinkers and brilliant communicators – we need 21st century skills, don’t you know. When, in truth it is memos, reports, box ticking paperwork and platitudes barely read or taken seriously.
Indeed, most of this hatred of AI comes from resentment, precisely because it reveals what we do to be so banal, throwing out emails, replying, tedious meetings, dull PowerPoints. Deep down we know that this should be short-circuited, automated, much of it even eliminated. Why write so much that remains unread? Why summarise meetings when most of them are unnecessary? Why run those courses when they have no measurable impact?
Many suppose that AI is an abstract thing, that can be tamed through debate, articles and papers. No, like water, or rather a rapidly rising tide that will not ebb, it has already infiltrated schools, universities, homes and workplaces. Law, healthcare, finance, retail – nothing is immune from its value. It’s a fifth columnist redefining things from within – because it is being USED, often on the sly.
AI feels powerful. That’s because it gives us agency, whether student, employee, applicant for a job, customer, patient… we suddenly have expertise at hand. Unsurprisingly, experts don’t like this, as that is what they sell.
Conclusion
The report has a more technical bent than most as the data was gathered from 730 Tech folk: developers, data teams, leadership, and others across technical roles and industries This makes it a bit more ‘real’ in my view, as many will be practitioners, people who not only use but have implemented AI in their organisations. Oddly, this also means it’s a bit out of date, as the data is from April, and we’ve had model updates and releases since then. Nevertheless, it is useful on some technical topics, such as what Vector Databases are being used.
No comments:
Post a Comment