Google ReCaptcha 3 always returning score of 0.9

Question

I am trying to guard against bot registration on my website by using Google ReCaptcha 3, I want to set a score threshold limit based on data collected over a period of time. However, the score is always coming back as 0.9.

I have set up a python script using selenium to constantly register to my site. I hoped that this would decrease the score over time however it is constantly staying at 0.9?

How can I reduce the score? Surely a script that is signing up 25 times a minute is enough to lower the ReCaptcha score?

I thought the whole idea of googles ReCaptcha v3 was to guard against bots in a discrete way?

def main():
browser = webdriver.Chrome(ROUTE TO CHROME DRIVER)

chrome_options = webdriver.ChromeOptions()
chrome_options.add_experimental_option("prefs", {
    "profile.default_content_setting_values.notifications": 2
})
chrome_options.add_argument("--incognito")
browser.get(MY SITE)
count = 50
while count < 100:
    browser.find_element_by_xpath('//*[@id="headerSignIn"]').click()
    browser.find_element_by_xpath('/html/body/div[2]/div/div[3]/div[1]/div[2]/p/a').click()
    browser.find_element_by_xpath('//*[@id="form_firstname"]').send_keys('Mr Robot')
    browser.find_element_by_xpath('//*[@id="form_email"]').send_keys('MrRobot' + str(count) + '@robot.com')
    browser.find_element_by_xpath('//*[@id="form_password"]').send_keys('[email protected]')
    time.sleep(2)
    browser.find_element_by_xpath('//*[@id="form_submit"]').click()
    browser.find_element_by_xpath('/html/body/div[1]/div/header/div[1]/ul/li[2]/a').click()

    count += 1

Nurse

Update the question with your code trials. – Advisedly

maybe it knows that a human is launching the script to test their own website. – Disapprobation

but all kidding aside... the google site seems to imply that placing the reCatptha on multiple pages helps it to gather data about behavior that is human as opposed to behavior that is bot-driven. They don't really disclose much about the score algorithm, but I'd think it would get better over time when they pull more data from your site. Sort of like a computer playing "Go" and giving percentages to the probably of the next move based on existing human-move data... – Disapprobation

I've had the same experience. I wrote a short Selenium script which tries to submit a form I'd like to guard, and was stunned to find that ReCaptcha invariably passes it with a 0.9 score. I understand that ReCaptcha is designed to improve as it processes more user activity over time, but I thought it ought to be able to handle something like my dumb-as-nails Selenium robot out of the box. – Rolfston

Same situation for me. Did you find a solution? I have reCAPTCHA v3 implemented in a form. I have filled in the form about 50 times. Half in an automated way with Selenium and the other half by filling in the form manually. In all cases, the score was 0.9 – Craps

Guys how is recaptcha enterprise ? – Cabala

have you tried without incognito mode? – Highpitched

Have you all tried using proxies? My score changes most of the time when using proxies. It might give the same high score when done from your personal IP as it is already authenticated by google? I will give a try and update ASAP. – Teredo