The easiest way for a hobby chemist to make acetic anhydride is probably by pyrolysis of acetone producing ketene:
If you value your life, you would avoid building your own ketene generator. If you're a chemical engineer or know one, you could probably pull off building a safe one. Alternatively I heard you can buy ketene generators (sometimes they're called ketene lamps). (Even then, bees have said the most they got out of their setups was 100 mL per day of continuous operation.) Sure, you can use contain and ventilate the apparatus, but where are you going to release the fumes? Outside? Ketene is as toxic as HCN. It's more responsible to just fucking steal AA than do this.
AA can be made starting with sodium ethoxide (from NaOH and anhydrous EtOH (dehydration of aqueous ethanol (190 proof vodka or cheaper less pure "denatured alcohol") with anhydrous CaO)) and disulfur dichloride (from gaseous chlorine (from hypochlorite salts and hydrochloric acid; or from electrolysis of sodium chloride solution) and elemental sulfur).
Do you generally tend to hysteria or are you somehow traumatized with ketene?
A little aspirator vacuum solves all problems of leaks and excess keten is washed down the drain for good. It dissolves reasonable enough in water to pose no problem in such a setup.
Albeit leaks should be no problem if you have just a little experience and some overpressure in the system is good for yields and furthers the reaction of keten with GAA which is astonishing slow as is the reaction of acetic anhydride with water btw. too. Excess keten is burnt off, at the end of the line comes a flameback arrestor and then a open flame and thats that. It will always burn for the methane but why not.
There have been many different incarnations of devices one can call "ketene lamp" and most worked some even good. The last stage of evolution used a SS heating element from a water heater or like this (courtesy lugh for providing the article and information) what eliminated the main problem of sooth formation almost completely (what was the main issue with the copper tubes).
The task is technically not trivial but nowadays with PID controllers costing 10 bucks and SSRs up to nothing, it is far easier then it was 8 or 10 years ago.
With the opportunity to present planned construction and to get quality feedback here or even elsewhere - SCM if you want to avoid the druggie thing a bit - its also pretty safe if one has a rest of reason left in ones brain.
Risky are other things, like crossing the main street on rush hour. Thats risky.
(For you have much more control on such a project then on mad drivers, yes?)
It is just that people nowadays are even less willing to invest real time and real money into something like this for they expect the manna to fall from heaven if they bark at the moon long enough.
And that why not so much informtion is thrown out onto the slightest question, because thats got annoying and boring and endless frustrating.
For me at least.
I though still enjoy the practice of installations which make a certain kind of people hysteric. Not because of this but because I love this technical tinkering....