We see this in all aspects of IT / IS industry. Companies want to test-drive data science, so they opt for one of two ways:
1) post job ad for a data scientist, but call it "Python Developer" or "Data Analyst".. and then lump 3 job roles into it (DBA, Analyst, Scientist), and offer them $50k
2) tell some tech person at work to go "learn data science".. because they think it's something you can just learn overnight, they won't have to spend any money to get the ball rolling on it, and they can use you to test-drive it
Both methods fail miserably...
1) DBA, Analyst & Scientist are 3 different roles each needing a $50k+ professional. By asking 1 person to show up to do all 3 roles, the person will be over-taxed managing data and trying to do analysis, they will never have time for research and exploration (the data science part). A competent Data Scientist will want $100k+, so they won't get a real data scientist. Instead, they'll get a faux-DS person.. someone that can stuff data through python algorithms, but doesn't really understand what's going on or why they're doing it. But, person has a job (probably desperate to get one right out of college).
2) person told to "go learn this" is basically chucked into the deep end of the pool with statistics and other high level stuff they have no clue about, and perhaps no interest in learning. But, they need to learn those high-level fundamentals to understand data science. This person will feel overwhelmed.
Both paths lead to a lackluster end result. Management will expect someone to just wave a magic data science wand and solve multi-million (billion) dollar problems with this fancy AI, Machine Learning, etc that managers have heard all about (through buzzwords, not real education). The person trying to do it will crash hard, and management won't see it as worth pursuring.. so they'll look for the next buzzword craze to go pursue so they can try to make it work and make it look good at their company and on their resume.
This is how many projects at companies (where management has no real idea what it entails) ends up flopping. Someone heard something, but has no clue about it. So, they want to do it, but don't want to invest a lot of time and money into it in case it flops (business managers invest in risk management.. and many of them think that involves mitigating risk by cheaping out on initiatives, so if it flops, it won't be that bad of a loss... but, they don't realize that by cheaping out they've hamstrung the initiative from the start, creating the inevitable blowup they worry about to begin with).
I had jobs do this to me in the past. If they were willing to foot the bill on learning / training.. ok, fine. But, if it was somethign radically different from my job description, I'd expect a pay raise to go with it. (EG: you're a web dev... them expecting you to be the new faux data scientist that will do AI magic at the company.. that's a different job description).
If they're not going to foot the bill on it, then I'd tell them that they don't understand what all is entailed in learning it.
To make it make sense to them, I often use learning a new language and culture as an example.
EG: I had a Project Manager come to me one time and tell me that I (a DBA) needed to field a web dev project. I had never done web dev, so I had no clue how to roll something like that out. They saw that I was generating some basic HTML pages, so they just assumed I was a web dev as well. But, what they wanted was active server pages and what-not.. things that would require a web server and lots more skill and resources then I had available to me.
To make the project manager understand what they were asking of me, I told them it would be like them going to China to become an HR person. They would need to learn the language (Chinese), and then how to do HR. It was two things totally out of their wheelhouse. But, they were asking me to just magically know how to do web dev overnight to roll out an online project tracking for a major system.
An alternate way to do this .... show them a bill.
Research all the classes and such you'd need to take to learn AI. Then write up an invoice or what-not and hand it to your boss.
When your boss goes "what's this?" You go, that's the cost estimate for training me on AI. Include the hours it would take you to learn it as well, because that's going to be the amount of down time you're going to need to learn the stuff where you won't be doing your normal job as well.
Because that's really what companies listen to.. time and money.. and time is money, so both translate into money.
Basically, you can balk about this.. or you can do like a lot of contractors do.. just nod your head and smile, then draft up a cost estimate that will show them that what they are asking is outlandishouly expensive. Once they see the time and money it will cost them, they'll rethink the strategy, b/c it adds context to the request in a way they can understand.
As it is right now, they don't understand anything about AI.. just that it's "an IT thing.. and we have an IT guy ... so tell him to do it."
Anything is possible. You just have to interpret it into time and money to make them understand if they really want to have you pursue it or not.
And, (as you said in repying to others) they are asking you to do all this in your spare time.. that is a sure-fire exclamation point that they are looking to explore data science for the absolutely cheapest cost possible.. IE: no cost to them, all cost to you.
That is not the kind of place you want to work at. Or, at least, you need to re-orient them to make them understand how outlandish their request is.
Again, submit a cost estimate, and when they balk jsut say "well, let's assume you were told to move to China and become an HR person.. but you were told to foot the expense on your own.."
You have to put it in their language for them to understand.. and that's where being an IT person and IT manager comes in. An IT manager knows how to translate stuff into businss terms to make other managers understand how ludicrous their requests are.
(As others have said, data science is an interesting field to pursue. But, right now it's all a hot mess. You have tons of foreign exchange grads coming out of college lying on their resumes about what they can do and how many years of experience they have all so they can get analyst jobs. You have companies that have no clue what data science is, so they put out job offers for trifecta DBA / Analyst / DS for $40k/yr without realizing how insulting that is. And, every company thinks that "real data scientists" eat, sleep and breath code, so they want you to have an extensive portfolio you can show them of work you've done on your own.. which is nuts, because it means you busted your hump doing a lot of work for no pay. Or, they want to see what you've done at other companies without realizing that would violate corporate espinage or IP theft laws. It's just ludicrous right now).
Source: I'm 15 years in the Data Analytics industry and just got a Masters in IS with heavy data science. I've been dealing with this stuff for a while now, as well as clueless management that doesn't realize how insane their requests are.