Should I Learn MS Access in the AI Era? (3 Viewers)

[SachAccess]

I am thinking about learning MS Access, but I have doubts because of the rise of AI. AI can now write complex queries, process data instantly, and provide high-speed results based on simple prompts. Given this, I’m wondering—is it still worth learning MS Access?

My concern is that by the time I become skilled, AI will be 10 times better and faster than me. I don’t want to give up learning due to fear of AI, but I also don’t want to invest time in something that might become outdated.

How do I convince myself that learning MS Access is still valuable? Have experts here faced similar doubts? Would love to hear real experiences rather than just relying on AI-generated answers.

Looking forward to your insights! Please help with your valuable inputs if you get time.

 

[Steve R.]

Well, it depends on what you want to do. The most obvious point, it is always good to learn new stuff like MS Access. The next consideration, is MS Access the database system that you would want to use. So you need to go back to think about the database solution that would be best for you. Since you appear concerned about implementing the "best" technology, I would suggest looking at an opensource database such as MariaDB along with Apache. As this is an MS Access forum, you will get a lot of very good MS Access advice.

 

[GPGeorge]

This question is hard to address because it's based on a false premise. That is, you seem to think that because the Large Language Models (LLMs) behind AI can produce potentially usable code rapidly, there's no reason for you to even try to use Access yourself. Nothing could be further from the truth. It's all the more reason you should invest time learning how to evaluate and modify and implement that potentially usable code appropriately.

What you bring to the table is not merely lines of code. What you bring to the table is the ability to envision how an Access database could be used to help solve a problem. You bring the ability to understand that problem in the first place. You bring the ability to propose approaches, designs and tools that might be brought to bear on the problem. And finally, you have the ability to formulate the right questions to pose for the LLM to research on your behalf.

AI cannot replace you; it can be a tool to make you more effective.

 

[The_Doc_Man]

We have seen many times that AI, using large language model as its base, produces erroneous code. Without a reasonable knowledge of Access, any VBA code and any designs that an AI produces will be suspect. We have seen cases where VBA gets conflated with VB - which DOES have a few differences from VBA. We have seen outright incorrect code. We have seen many cases on this forum where code produced by an AI simply will not run because it doesn't understand the boundaries between different products and also between vendor code vs. privately developed code. Until we can get AI to the point that it doesn't hallucinate (and no, that's not a dig; it is the industry name for AI "flights of fancy"), it cannot be trusted to generate proper code.

Therefore, while AI might one day reach that point where we CAN trust it to write code, it is not there yet and I don't see it getting there in the current direction of "better training". Inherent in the LLM is the fact that it gives the statistically best answer - but if a lot of the training answers were based on a generally incorrect notion, then statistically, the odds are that the answer was wrong.

As to whether you should learn Access and VBA, that choice depends on what you intend to use for DB development. Whatever you DO choose, study that. IF you choose Access and VBA, then study what you choose to use. If you choose to use ALGOL from the 1960s, study that. It is never wrong to learn about the tools you choose to use.

 

[CJ_London]

and provide high-speed results based on simple prompts

To get the right answer you need to be able to ask the right question and provide all the relevant parameters which could impact the answer. That requires training.

When you develop an app in access/web/ whatever you are the AI - you need to ask questions of the client to clarify exactly what the requirements are - and often you discover ‘wrinkles’ the client hasn’t considered so further clarification is required

When AI gets to the stage of understanding the resources available and can ask relevant clarification questions perhaps it will replace programming per se but in the meantime learning about managing and presenting data using access/web/whatever gives you a good grounding in how to ask questions and identify relevant parameters

 

[Mike Krailo]

MS Access is not that hard to learn. It's one of those programs that just won't go away because it's so good at doing what it does which is make it much easier to develop a fully working business solution in record time compared to other methods. When coupled with SQL server there isn't much that can limit what you can do with it. The question about AI being able to answer questions about building queries is irrelevant in terms of replacing Access because Access is more about the form and report objects that make up the user interface. What does AI do to help you develop a useable interface quickly?

 

[GPGeorge]

Spoiler:

MS Access is not that hard to learn. It's one of those programs that just won't go away because it's so good at doing what it does which is make it much easier to develop a fully working business solution in record time compared to other methods. When coupled with SQL server there isn't much that can limit what you can do with it. The question about AI being able to answer questions about building queries is irrelevant in terms of replacing Access because Access is more about the form and report objects that make up the user interface.

What does AI do to help you develop a useable interface quickly?

I think the operative phrase in an answer to this question could be "Fail Fast." With an interface design, that's less apparent, of course. You have to know which battles to fight. LLMs may or may not give us much assistance with Interface Design. I don't know. I might see if I can figure out how to do that.

On the other hand, as others have pointed out, potentially usable code from an LLM can be tried quickly, and if it's not valid, you'll fail fast. That means you can move on fast.

The iterations needed to get to minimally usable code won't differ much from hand-written code, at least for a less experienced developer. However, with the assistance of an LLM, that developer can cycle through multiple iterations more rapidly, and even with more variations, than that same less experienced developer can think up and try.

As I said before, I think the entire premise underlying the question is inappropriate. It's not a question of "AI" versus "HI". It's a question of HI learning how to incorporate AI into our development.

As a matter of fact, I've turned 180 degrees in my thinking on this. My early attempts to get usable code from ChatGPT soured me on its ability. Only after time and further reflection did it dawn on me that I've written plenty of faulty trial-and-error code and SQL over the years. For years, I prided myself on having a "try it and see what happens" attitude. Well, it turns out that LLMs can make it much easier and quicker to "try it and see what happens".

 

[GPGeorge]

We have seen many times that AI, using large language model as its base, produces erroneous code. Without a reasonable knowledge of Access, any VBA code and any designs that an AI produces will be suspect. We have seen cases where VBA gets conflated with VB - which DOES have a few differences from VBA. We have seen outright incorrect code. We have seen many cases on this forum where code produced by an AI simply will not run because it doesn't understand the boundaries between different products and also between vendor code vs. privately developed code. Until we can get AI to the point that it doesn't hallucinate (and no, that's not a dig; it is the industry name for AI "flights of fancy"), it cannot be trusted to generate proper code.

Therefore, while AI might one day reach that point where we CAN trust it to write code, it is not there yet and I don't see it getting there in the current direction of "better training". Inherent in the LLM is the fact that it gives the statistically best answer - but if a lot of the training answers were based on a generally incorrect notion, then statistically, the odds are that the answer was wrong.

As to whether you should learn Access and VBA, that choice depends on what you intend to use for DB development. Whatever you DO choose, study that. IF you choose Access and VBA, then study what you choose to use. If you choose to use ALGOL from the 1960s, study that. It is never wrong to learn about the tools you choose to use.

I think the way to approach AI is not as an either/or choice between "AI" and "HI". And I don't think the production of hallucinated responses is a deal killer.

To reach for a tangentially relevant historical analogy, I refer to the famous Thomas Edison quote about creating the first light bulb. "I have not failed 100 times, I found 100 ways that do not work".

The point is that development is an iterative process. And that means we should not be afraid to "try it and see what happens" as many times as it takes to get it right.

We can't trust LLMs to generate proper code? So what? AIs can produce volumes of sample code quickly and iteratively. It's up to us to "try that sample code and see what happens"! One alternative to that, of course, is searching sites like this one to see if someone posted code similar to what we want.... I'd rather let an LLM do that searching across this site and many, many other sites at a much faster pace.

 

[SachAccess]

Thanks a lot everyone. Apologies for late reply. Please give me some time to read all the replies and respond. Have a nice day ahead.

 

[AccessBlaster]

As with any tool, AI has a learning curve understanding how to prompt correctly, verifying results, and refining your approach. Your vision for your application will evolve over time, requiring different methods to achieve your goals.

For example, you initially planned for 10 local users, but your boss was so thrilled by your success that he now wants to scale it internationally. Now what? You're in trouble because you only had time to learn MS Access.

However, if you have a paid AI solution integrated into your application, you can now prompt it for the best approach whether that involves .NET, C#, SQL, or other technologies like cloud or API.

To be clear, I’m not criticizing MS Access. I just see AI as a potential workaround for complex challenges, it’s still your vision.

 

[Mike Krailo]

As I said before, I think the entire premise underlying the question is inappropriate. It's not a question of "AI" versus "HI". It's a question of HI learning how to incorporate AI into our development.

This is true. Well said. I'm pretty sure those DOGE kids are using AI very effectively to accomplish the task at hand. We all need to get at using it to accomplish our own tasks at hand.

 

[SachAccess]

First, I forgot to mention that the entire post of mine is based on my perception and which could be entirely wrong.
Hence I requested help and getting immense help.
I work in Excel, I provide prompt to GPT and most of the times get good response. But I forgot that it might be quality of my prompts which is giving me useable code. If I need to repeat same prompt for Access then I must know Access so right prompts will be asked.
I asked a similar question to GPT about preference to Python and Access, which is safe to learn considering AI. GPT gave me answer as Python, to which I as not convinced. My belief is, Forms, Report and GUI would require manual efforts. Other real world scenario in my limited experience I have seen that, companies prefer free or MS products only. Companies do not prefer paid platforms. My experience is in BPO sector.

 

[SachAccess]

Well, it depends on what you want to do. The most obvious point, it is always good to learn new stuff like MS Access. The next consideration, is MS Access the database system that you would want to use. So you need to go back to think about the database solution that would be best for you. Since you appear concerned about implementing the "best" technology, I would suggest looking at an opensource database such as MariaDB along with Apache. As this is an MS Access forum, you will get a lot of very good MS Access advice.

I will be honest here. I have 1 short term goal and 1 long term goal. Might look like a wishful thinking too.
Short team goal is, to learn MS Access, the only reason here is I need to safeguard myself from being obsolete.
My main concern, is the short term goal really that safe or just my assumption.
One reference, Excel + Access combination will be picked by HR quickly rather than Excel + Python combination.
Don't know good word for bias but that is what I feel.

 

[SachAccess]

This question is hard to address because it's based on a false premise. That is, you seem to think that because the Large Language Models (LLMs) behind AI can produce potentially usable code rapidly, there's no reason for you to even try to use Access yourself.

You are absolutely correct. Thanks a lot for being so honest and clear with your guidance.

 

[Edgar_]

Short team goal is, to learn MS Access, the only reason here is I need to safeguard myself from being obsolete.

One reference, Excel + Access combination will be picked by HR quickly rather than Excel + Python combination.

I am thinking about learning MS Access, but I have doubts because of the rise of AI. AI can now write complex queries, process data instantly, and provide high-speed results based on simple prompts. Given this, I’m wondering—is it still worth learning MS Access?

Hear me out. The forum owner recently mentioned that traffic has increased, but as someone who enjoys a good challenge, I haven't seen that many challenges here lately. Could it be that more people have adopted Access thanks to the AI? And therefore, since the AI helps solve issues, they don't need to post?

I don't know.

Maybe it’s time you start using Access. But I recommend checking the latest Stack Overflow survey, where they list the most used technologies. The MS Access/VBA stack usually ranks around 30th, and that’s been consistent. Keep an eye on the 2025 survey coming out in a few months, if it drops lower, it might be worth considering another tech stack.