AI Math Tools Help With Dyscalculia When Your Child Still Does the Thinking

Do AI math tools help a child who finds math hard?

Here is the honest answer from the math research we hold. Benavides-Varela and colleagues (2020) reviewed trials of math programs on a computer. On average, the programs helped. But the results swung widely from one study to the next. The authors said that spread was too wide for strong conclusions.

One finding stood out. Math games did no better than plain practice programs. The review found “no evidence that videogames offer additional advantages.” To us, that says the flash is not the part that helps.

A second study, by Kohn and colleagues (2020), tested computer training. The children in it struggled with math. The trained children gained more in arithmetic. They also got better at placing numbers on a number line. Children who felt less worried about math gained more. So did children who did not also find reading hard.

These studies tested math software: drill, tutoring programs and games. The newer chat-style AI is a different tool, so ask it the same questions. The word “revolution” oversells all of this. A good tool helps with plain, steady practice. That is worth having.

What kind of practice moves math forward?

Haberstroh and Schulte-Körne (2019) summed up the German clinical guidelines on math difficulty. Practice on the actual math helped more than no help at all. It also helped more than training a separate skill, like working memory. The guideline favored one-to-one work. It also says strong trials are still missing.

So the first question for any app is simple. Does it practice the math your child is stuck on? That might be what a number means. It might be how two numbers compare. Our view: that is the work a good tool should do.

Worry is a separate question. Devine and colleagues (2018) found that math difficulty and math anxiety are largely separate. In some children, they show up together. So it helps to know which one you are seeing. Some children need help with both.

You do not need a label to start. Kißler and Kuhn (2025) found a fluid line between dyscalculia and ordinary math struggle. Their advice is to support every child who finds math hard. Mazzocco and Räsänen (2013) add that math difficulty is not fixed year to year. Teaching is part of the picture.

Math trouble also tends to be spotted late. Jordan and Devlin (2023) warn that it is easy to miss for years. Price and Ansari (2013) say the field has had little research to build on. That gap belongs to the system. Here is why math struggles get caught later than reading struggles.

How do you tell if a tool builds the skill or does the math?

Ask one question of any tool. Is this support building the skill, or replacing the expectation that it gets built? That is our test. It is not a vote against supports, an IEP or a 504 plan. It is a way to judge one tool.

Stefan Bard makes a related point in the video on this page. He advises schools on AI. His advice: use AI to make yourself better, not to do your work. For a child who finds math hard, that means the child does the thinking. The tool sets up the problem and checks the answer.

Sit with the app for ten minutes and watch for these signs:

• Your child works out each answer. The tool does not show it first.
• The practice is the math your child is stuck on, not a side skill.
• When your child gets one wrong, the next problem changes to match.
• Points and badges are extras, not the main event.
• You stay close for part of it and talk about the numbers.

Then check the settings tonight. Does the app give the answer after one miss? Look for a way to turn that off. A hint that points to the next step is fine. A full answer handed over is the shortcut to watch for.

A points system makes a child feel like they’re learning. Real mastery makes them feel capable. Our view: a tool works best with you nearby, not in place of you. Here is one question to ask about a school’s AI program.