AI Music Creation Platform that can Make a Billion Songs

    Published on:

    According to a new study, researchers at the University of California have found that AI bots can successfully defeat CAPTCHAs (security measures used to separate humans from bots online), even better than humans and far more powerful. found to be faster than

    preprint of research paper A recent paper published on arXiv found that AI programs were able to solve CAPTCHAs 15% more accurately than humans. To do this, the bot mimicked how the human brain and vision work.

    As a result, the system can now perform many of the tasks once thought to be the domain of humans, researchers say. Scientists fear this could make websites that rely on the technology vulnerable to spam and other malicious activity.

    Also read: AI cyberattack steals passwords with 95% accuracy, study warns

    Smarter AI bots

    The research was conducted by a team of researchers including three from the University of California, Irvine, and one each from ETH Zurich, Microsoft, and Lawrence Livermore National Laboratory. The title is “Empirical research and evaluation of his CAPTCHA in the present age”.

    “Bot accuracy ranges from 85% to 100%, with the majority above 96%, which is well above the human accuracy range we have observed (50% to 85%),” the research paper states. is listed.

    “Furthermore, the bot resolution time is significantly reduced in all cases except reCAPTCHA, which is about the same as the human resolution time of 18 seconds and the bot resolution time of 17.5 seconds.”

    A Completely Automated Public Turing Test to Tell Computers and Humans Apart (CAPTCHA) is a technology used to protect websites from fraud and abuse. CAPTCHAs typically present users with problems that are easy for humans to solve, but difficult for bots.

    Examples of CAPTCHAs used in research.

    For example, a CAPTCHA might ask the user to identify distorted text, select a rectangle with a car, or solve a simple math problem. The technology is designed to effectively prevent AI bots from accessing websites, but it can also be frustrating when used by humans.

    In 2019, Google replaced CAPTCHA with an advanced version called reCAPTCHA. It is designed to be difficult for bots to decipher.

    However, Aaron Malenfant, the technical leader of the team, said that this technology will not be possible in 10 years because the technology that can run the Turing test in the background like Quartz has improved. report.

    “Unloved” capture

    For the study, researchers selected over 1,000 people to test websites that use CAPTCHA challenges. These websites account for 120 of the world’s 200 most popular websites ranked on the Alexa Top Websites list.

    Participants were asked to solve 10 different CAPTCHAs, including identifying boats and chimneys, rotating images, checking checkboxes, and entering distorted text.The study was conducted on Amazon’s crowdsourcing platform MTurk.

    Research papers show that humans take longer to solve CAPTCHA puzzles than AI programs when placed in more natural environments. In such an environment, a human would complete a puzzle in 22 seconds, while an AI bot’s average solution time would be 17.5 seconds.

    “We know for sure that they [the tests] So unloved We didn’t have to do any research to come to that conclusion,” says one of the researchers, Gene Tudick. Said new scientist.

    “But people don’t know if the effort, the enormous global effort that is being invested in solving CAPTCHAs every day, year after year, month after month, is really worth it.”

    Can digital identities limit Web3 threats?

    Experts say that as AI becomes more sophisticated, it will become increasingly difficult to distinguish between real and fake identities. As CAPTCHA research shows, it could already be happening.

    Artificial intelligence can also compromise the security and privacy of digital identities. As an example, it can be used to create deepfakes, i.e., real but fake images and videos that are used to impersonate another person, including someone else’s voice.

    Some companies such as Polygon, Identity Labs, and Worldcoin are currently trying to combat these threats by building blockchain-based digital identities. Worldcoin is developing a World ID system that will scan a person’s iris to collect unique biometric data.

    The recently launched Identity Labs NFIDsa passwordless decentralized identity and login tool that allows users to verify their identity by linking their phone number to their account.

    Experts say digital identity is the cornerstone of Web3 and the Metaverse, as it enables trust and security in decentralized systems. Digital IDs come in two forms. The first is a digital version of your official physical identification, such as your passport, stored in your mobile cryptocurrency wallet.

    The second is credentials for accessing online services such as DeFi apps, NFT marketplaces, and other Web3 services. In both cases, we use digital IDs to verify your identity and ensure that you have the necessary permissions to access our Services or perform certain actions.


    Leave a Reply

    Please enter your comment!
    Please enter your name here