Head to head · October 2026 cycle
Phrasly vs WriteHuman
WriteHuman finished 13.1 points ahead of Phrasly in the October 2026 cycle, 74.89 to 61.79, ranking #1 against #12 in a field of 14. WriteHuman scored higher on 3 of 6 detectors and 6 of 6 writing categories. Phrasly came out ahead on Pangram (85.6 vs 50.3), Copyleaks (100.0 vs 93.3) and bypass rate (95.3 vs 88.8). The bypass-rate gap, 95.3 to 88.8, sits inside both tools' 95% confidence intervals, so treat bypass as a draw.
- Detectors
- 3 – 3
- Categories
- 0 – 6
- Components
- 2 – 3
30 prompts · Last tested · Methodology v1.3.0
Score components
What the overall score is made of: bypass rate weighs 42%, meaning preservation 32%, readability 16%, consistency across categories 10%. Output-quality penalties come off the total. How scoring works →
CIs overlap marks a gap that sits inside both tools' published 95% confidence intervals for that component; it may not survive another cycle.
Detector by detector
Share of each tool's outputs that a detector classified as human-written. Split 3–3.
Writing categories
Category score per writing context: bypass credit only for real rewrites, blended with the category's own meaning preservation and readability. Small categories swing; the prompt count is shown on each row. How category scores work →
Head-to-head history
Overall score in every cycle both tools were tested. WriteHuman has finished ahead in 5 of 5 shared cycles.
| Cycle | Phrasly | WriteHuman | Margin |
|---|---|---|---|
| October 2026 | 61.79 #12 | 74.89 #1 | → 13.1 |
| September 2026 | 67.58 #8 | 78.29 #1 | → 10.7 |
| August 2026 | 63.30 #7 | 76.69 #1 | → 13.4 |
| July 2026 | 61.47 #9 | 73.07 #1 | → 11.6 |
| June 2026 | 71.10 #8 | 83.59 #2 | → 12.5 |
Where each one wins
Every measure above that one tool won outright, largest margin first within each group.
Phrasly
5 wins- Bypass rate Score 95.3 vs 88.8
- Consistency Score 96.9 vs 94.7
- Pangram Detector 85.6 vs 50.3
- Copyleaks Detector 100.0 vs 93.3
- ZeroGPT Detector 94.3 vs 89.5
WriteHuman
12 wins- Penalties Score None vs −11.0
- Readability Score 55.9 vs 44.2
- Meaning preservation Score 60.0 vs 49.9
- Originality.ai Detector 100.0 vs 96.6
- Winston AI Detector 100.0 vs 97.0
- GPTZero Detector 99.5 vs 98.6
- News Article Category 77.9 vs 16.8
- Marketing Copy Category 71.8 vs 45.8
- Discussion Board Category 71.9 vs 52.9
- Blog Post Category 72.4 vs 57.8
- and 2 more
Same prompt, both outputs
Both tools were given this AI-written passage in the October 2026 cycle. Below it is what each one returned. This prompt was picked by a fixed rule, not by how the tools scored.
AI-written input
Every tool on this prompt →News Article · News Article · written by claude-sonnet-5-5 · 444 words
**Digital Privacy Laws Struggle to Keep Pace With Data-Hungry Technology**
As lawmakers worldwide celebrate landmark privacy regulations, researchers and advocates warn that the laws are straining against the realities of modern data collection, leaving significant gaps in how personal information is protected.
Europe's General Data Protection Regulation, California's Consumer Privacy Act, and a growing patchwork of state and national laws have given individuals new rights to access, delete, and limit the use of their data. Yet critics say these protections often prove more symbolic than practical.
One central problem is consent. Most privacy frameworks rely on the idea that users agree to data practices after reading disclosures. In reality, few people read lengthy privacy policies, and fewer still understand them. A frequently cited study estimated that reading every policy an average person encounters in a year would take weeks of working time. Consent banners, meanwhile, are often designed to nudge users toward clicking "accept."
Enforcement presents another obstacle. Regulators are chronically understaffed relative to the thousands of companies they oversee. Investigations can take years, and fines, while sometimes large in absolute terms, may amount to a small fraction of revenue for major technology firms. Some companies treat penalties as a cost of doing business.
The borders of the internet also complicate matters. Data flows across jurisdictions in milliseconds, but legal authority stops at national lines. Disputes over transferring European data to the United States have repeatedly invalidated agreements, creating uncertainty for businesses and users alike.
Technology itself is outrunning legal definitions. Privacy laws typically protect "personal information," but advances in data analysis allow companies to re-identify supposedly anonymous records by combining datasets. Inferences drawn by algorithms, such as predictions about health, finances, or political leanings, occupy murky legal territory because they are generated rather than provided by the user.
Meanwhile, large categories of data collection fall outside existing rules. Data brokers assemble detailed profiles from public records, app activity, and purchased information, often with little direct relationship to the people profiled. Government access to commercially held data raises further questions, as agencies can sometimes buy information they would otherwise need a warrant to obtain.
Privacy advocates argue for a shift away from individual consent toward limits on what companies may collect and do in the first place, an approach known as data minimization. Some proposals would ban certain practices outright, such as selling sensitive location data.
Industry groups counter that overly rigid rules could stifle innovation and burden small businesses. The debate is likely to intensify as artificial intelligence expands the appetite for training data.
For now, experts say, individuals remain largely responsible for protecting themselves, a burden that few are equipped to carry.
Show the full input Show less
438 words (input 444)
As governments continue to develop more strict privacy laws researchers are beginning to show concern over whether these laws can effectively cover the amount of data being collected by companies. Many countries have developed privacy acts that allow their citizens to know what data companies are collecting and can delete or stop companies from using their data. Although there are flaws with these new regulations many people believe it to be a positive. There are many drawbacks to these laws. Privacy policies allow companies to gain your consent to use your data. Since most people do not read privacy policies its almost impossible to truly tell if someone knows their data is being used. It's been said that it would take the average person 7+ hours a day for one year just to read all the policies you'd come across. There are many reasons why people don't take the time to read these policies and some people even say the companies purposely do this to ensure consent. Another issue with privacy laws is enforcement. There aren't enough employees within regulatory agencies to keep up with all of these companies. When cases are investigated it can take multiple years. Once fines are given to companies it can sometimes seem cheap compared to their earnings. Because of how global our world has become with internet and data. There is so much gray area when it comes to international laws and regulations. Recently we've seen many examples of why moving EU data to the US isn't possible and will create issues. As our technology advances there is more gray area around privacy laws. What is considered personally identifiable information? Companies can use databases of information and figure out who people are by matching different sets of data together. Also with all the new information being derived from peoples data where does the line get drawn when you didn't give that information. Not all forms of collecting data will fall under these new laws. Companies such as data brokers buy and sell peoples information with no knowledge of that person. Governments also can use this data to find information about people that they wouldn't have been able to obtain without a warrant.
With more technology and companies developing ways to profit off peoples data. There are calls for change from those who believe that people should have more control over their information. One example would be limiting what information a company can collect on you.
Privacy laws are something that will continue to grow and evolve with our world. Currently its up to each of us to help secure our own information.
Show the full output Show less
- GPTZero1.00passed
- Pangram1.00passed
- Winston AI1.00passed
- ZeroGPT0.98passed
- Copyleaks1.00passed
- Originality.ai1.00passed
Passed 6 of 6 detectors.
Every Phrasly run in October 2026 →492 words (input 444)
New Privacy Laws Don't Keep Up With New Technology
Recently, new privacy laws have passed around the world and lawmakers have celebrated these achievements. However, new gaps have emerged in the legal framework protecting user's privacy.
In the last few years, laws like the GDPR and CCPA, have granted users rights over their personal information by giving users the ability to view or erase the information businesses collect about them. But, many agree that these changes have been mainly symbolic.
A major issue with enforcing new privacy laws is getting companies to obtain user's consent to collect and use their information. In order for users to truly consent, they need to have an understanding of the information they are agreeing to. However, given the amount of time it would take to read and understand all the terms and conditions a user would encounter in a year, and the fact that most users don't read these agreements, obtaining true consent is impractical.
Because of limited resources, regulators prioritize ensuring companies adhere to the new privacy laws by enforcing an adequate level of transparency. This has led to a surge in the use of black letter, or strictly legal, language in user's agreements.
With information moving so quickly across international boundaries and so many of the world's businesses operating globally, users and businesses remain confused about what they are allowed to do and what businesses are legally allowed to do.
Privacy laws are slow to keep up with changes in technology. Laws generally take a long time to pass, and once laws are on the books they usually don't keep up with changes in technology. What's more, companies often combine the personal information they've collected with other datasets in order to make inferences about people. Legal challenges often arise around inferences that are harmful, sensitive, and/or incriminating. But because these inferences are often made by companies and are not actively provided by an individual, they often fall into a legal gray area.
There are many ways that companies gather information that are outside the purview of the law. Data brokers collect information about people who are unrelated to the data that's collected, often by purchasing information and compiling detailed profiles. The government can purchase personal information as well.
Some argue that in order for people to have greater control over sensitive and personal information, the law should support restraints on what information companies are allowed to collect and use in the first place. There are proposals that provide outright bans on some practices, such as the ban on the sale of sensitive personal information.
Many companies are collecting increasingly larger amounts of information in order to train and develop artificial intelligence. There are few, if any, legal constraints on what information companies collect and/or how they use it. As such, it will fall on people to be able to decipher what companies are doing, and protect themselves, which most people are unable to do.
Show the full output Show less
- GPTZero1.00passed
- Pangram1.00passed
- Winston AI1.00passed
- ZeroGPT1.00passed
- Copyleaks1.00passed
- Originality.ai1.00passed
Passed 6 of 6 detectors.
Every WriteHuman run in October 2026 →Each chip is one detector's score for that output, from 0 (AI) to 1 (human). A score of 0.50 or higher counts as a pass and shows green; below that the detector caught it and the chip shows red.
Questions people ask
Is Phrasly better than WriteHuman?
WriteHuman finished 13.1 points ahead of Phrasly in the October 2026 cycle, 74.89 to 61.79, ranking #1 against #12 in a field of 14. WriteHuman scored higher on 3 of 6 detectors and 6 of 6 writing categories. Both tools ran the same 30 prompts, scored by the same 6 detectors, under methodology v1.3.0; every input, output and verdict is published.
Which bypasses AI detectors better, Phrasly or WriteHuman?
Phrasly posted the higher bypass rate, 95.3 vs 88.8, and scored higher on 3 of 6 detectors. Phrasly led on Copyleaks, ZeroGPT and Pangram; WriteHuman led on GPTZero, Originality.ai and Winston AI. The overall bypass gap is inside both 95% confidence intervals, so it is not a reliable difference.
Do Phrasly and WriteHuman keep the original meaning?
WriteHuman preserved meaning better, 60.0 vs 49.9 on our 0–100 similarity scale. On readability, WriteHuman rated higher, 55.9 vs 44.2. Penalties this cycle: Phrasly −11.0, WriteHuman None.