For decades, storing paper receipts in a shoebox was how we defined our annual tax filing process. For most of us, that chaotic, paper-chasing tradition might be finally coming to an end. AI-driven automation helps with simple, repetitive tasks such as data entry, data transfer and math.
AI-driven tax strategy, on the other hand, uses machine learning and natural language processing to identify tax deductions and share regulatory and compliance information.
That’s not fully here yet: An AI tool that can analyze complex business tax structures while advising on strategies doesn’t exist right now, said Bill Park, CEO of tax intelligence platform TaxHakr.
But that day is coming, as advances in AI and automated processes help taxpayers and tax professionals streamline processing and reduce tax liabilities.
Not everyone agrees that AI integration is the best path forward with tax preparation. Many critics have ethical reasons for avoiding its use. A study by the Santa Clara Markkula Center for Applied Ethics found that 82% of respondents care about the ethics of AI. CNET reported last year that AI experts feel optimistic about the positive impact of AI in the coming decades.
We break down the pros and cons and how AI’s powerful automation is transforming how we file our taxes.
The pros and cons of AI usage in tax preparation
Pros
- Potentially faster refunds
- Lower cost
- Reduced stress
- Larger refunds (since AI may spot deductions and credits humans miss)
- Lower audit risk (since AI can spot red flags)
- Reduction in errors
Cons
- Data privacy risks
- Potential ethical issues and algorithmic bias
- Fairness
- Limited transparency in AI decision-making
- Keeping a human in the loop adds time and costs
- Difficulty in fixing deep learning systems when they produce unwanted outcomes
AI and taxes: The benefit of speed and accuracy
AI can reduce errors, speed up processing of returns, and ensure businesses and individuals claim the deductions and credits they deserve. Here’s all the ways AI can work for you during tax season.
Error reduction
Manual data entry takes hours, as the average person types about 40 words per minute with 92% accuracy. In contrast, with automated AI filing, this task takes mere minutes, as optical character recognition technology, combined with natural language processing,scans typed or handwritten forms or receipts in seconds.
Cross-referencing and validation are manual processes that can take hours, requiring time and attention to verify numbers and compare them with prior years (adding approximately 20 minutes of manual work per form).
Consumer and small-business tax software, like TurboTax and H&R Block, and large-scale platforms tax firms use (such as CCH Axcess) combine OCR and NLP to read and categorize W-2s, 1099 forms, receipts, bank statements, and other financial paperwork.
AI automation reduces it to minutes by comparing figures and quickly recognizing patterns. Reviewing tax returns, which requires in-depth knowledge of the current tax code, takes hours of manual work. Automated tax review still needs human oversight — you’d still want a credentialed tax preparer to sign off on your tax return — but the streamlined process reduces the time from hours to minutes.
The overall manual process can take days to weeks because tax returns go through many hands. AI and automation can streamline it to hours, since AI enters and sorts data, verifies entries and identifies anomalies, leaving humans to review the results.
Automatic categorization
Machine learning models, combined with natural language processing, are used on tax platforms and in accounting software to categorize income and expenses based on transaction patterns and merchant details. The model learns based on prior inputs and user feedback.
Sorting business and personal expenses becomes especially important for 1099 contractors, gig workers, and other self-employed individuals who may not always keep perfect records.
AI systems can track expenses by category and use context to determine business deductions versus personal expenses. H&R Block’s Self-Employed package, highly recommended by CNET staff, takes filers through an interview process to identify self-employed deductions and also looks for industry-specific expenses to offset 1099 income and boost your tax refund.
“If you plug in data, most systems are going to give you general feedback,” Park said. “It potentially gets people thinking about things they may have missed and probably encourages them to ask their CPA some questions.”
Workflow efficiency
Global legal, tax, and news information provider Thomson Reuters predicts that AI can save tax professionals 5 hours per week, equivalent to $24,000 in annual value.
Cloud-based AI platform and developer Dodocs.ai says its current AI tax prep tools reduce data entry time by 95% while achieving 99.5% accuracy.
Compliance and regulation
AI-powered tax research software can deliver answers regarding the tax code in seconds, versus people who have to manually sift through pages of tax code, especially as rules changed with the introduction of the Trump administration’s One Big Beautiful Bill.
Even experienced accountants might need to look up details in the tax code due to its complexity and frequent changes, said Park.
The shift to advisory services
AI can help free up CPAs (certified public accountants) and other tax professionals from spending time on data entry and compliance, allowing them to focus on strategic planning and advisory services.
“In the bookkeeping portion of our business, we see 100% of our clients overpaying on taxes,” Park said.
AI automation for tax strategies can help small businesses, individuals and gig workers structure their tax filings to save money. It can assist with strategic planning. It might suggest ramping up FSA or HSA contributions, for example, or waiting to make a charitable donation until the new year.
Manual vs. automated filing: Time spent on key tasks
| Key task | Manual filing process | AI-powered filing process |
|---|---|---|
| Data entry | ~Hours | ~Minutes |
| Cross-referencing and validation | ~Hours | ~Minutes |
| Review | ~Hours | ~Minutes to hours |
| Overall Process | ~Days to weeks | ~Hours to days |
The downsides to AI and tax filing
Anytime we integrate AI into our personal lives, we will face new risks. Here are the risks when it comes to your taxes
Data security, privacy, and ethics
Tax prep is a highly personalized endeavor in which preparers handle sensitive personal and financial information.
While most AI tax software and tools follow the security and data privacy requirements of banks and fintech companies, there’s another concern: the use of personal data to train AI systems.
The 2022 White House Blueprint for an AI Bill of Rights recommended that AI developers seek consent for data use. The Trump administration has recently shifted government policy to reduce barriers for AI companies. That doesn’t necessarily preclude data privacy and consent, but it doesn’t necessarily require these protections, either.
Potential algorithmic bias
Algorithmic bias is another risk of AI systems. In 2023, the Stanford Institute for Economic Policy Research found that Black taxpayers are three to five times more likely to be audited than other groups.
AI algorithms base their decisions on past inputs, spotting patterns and repeating them. If IRS AI software follows past patterns that have been proven biased, it could have costly consequences for taxpayers.
AI’s ‘black box’ problem
Technology experts recognize a “black box” problem inherent in AI use. Humans have no insight into exactly how AI makes complex decisions, any more than we understand how the human brain collects data, quickly sifts through patterns and draws conclusions.
“This quality makes it difficult to fix deep learning systems when they produce unwanted outcomes,” writes Samir Rawashdeh, an associate professor at the University of Michigan-Dearborn.
It also makes it difficult to know if the AI model is capable of making the correct conclusions most of the time.
If tax preparers can’t trust the AI model, the technology becomes a roadblock rather than a loyal, time-saving assistant. Testing can reveal bugs, but it can’t reveal holes in the system. Like humans, AI doesn’t know what it doesn’t know.
Human oversight
Keeping a human in the loop is the best way to spot errors before they become a liability or an audit risk.
“Automated tax filing is cool, but if you’re automating bad information, you’re just doing it faster without better results,” Park said.
Human CPAs or tax attorneys can interpret ambiguous tax codes and analyze complex scenarios to maximize deductions while minimizing audit risk.
The future of AI and taxes
The future of tax prep in this AI world isn’t about AI replacing humans. Instead, AI use can be a powerful, beneficial partnership between the two.
AI can extract data from financial forms, receipts and tax forms, increasing the speed and accuracy for tax preparers and DIY filers. However, human oversight remains at the forefront of tax preparation, tax avoidance strategies and audits.
AI is a powerful tool that helps improve research and accuracy, but tax professionals have decades of experience and training that is difficult to replace.
The most exciting part about AI and taxes is how it brings tax advice access to middle-class and low-income workers, said Katie Collins, content director for taxes at CNET. “Before AI, sophisticated tax strategies were provided to those who could afford to pay for it, which helped create a fair, more balanced tax system.”
During our testing, we found that H&R Block has the best use of AI and the best overall at-home tax software.
