Phishing is the fraudulent attempt to obtain sensitive information such as usernames, passwords and credit card details by disguising oneself as a trustworthy entity in an electronic communication.^[1][2] Typically carried out by email spoofing^[3] or instant messaging,^[4] it often directs users to enter personal information at a fake website which matches the look and feel of the legitimate site.^[5]

Phishing is an example of social engineering techniques being used to deceive users. Users are often lured by communications purporting to be from trusted parties such as social web sites, auction sites, banks, online payment processors or IT administrators.^[6]

Attempts to deal with phishing incidents include legislation, user training, public awareness, and technical security measures (the latter being due to phishing attacks frequently exploiting weaknesses in current web security).^[7]

The word itself is a neologism created as a homophone of fishing.

Phishing attempts directed at specific individuals or companies have been termed spear phishing.^[8] In contrast to bulk phishing, spear phishing attackers often gather and use personal information about their target to increase their probability of success.^{[9][10][11][12]}

Threat Group-4127 (Fancy Bear) used spear phishing tactics to target email accounts linked to Hillary Clinton's 2016 presidential campaign. They attacked more than 1,800 Google accounts and implemented the accounts-google.com domain to threaten targeted users.^[13][14]

The term whaling refers to spear phishing attacks directed specifically at senior executives and other high-profile targets.^[15] In these cases, the content will be crafted to target an upper manager and the person's role in the company. The content of a whaling attack email may be an executive issue such as a subpoena or customer complaint.^[16]

Clone phishing is a type of phishing attack whereby a legitimate, and previously delivered, email containing an attachment or link has had its content and recipient address(es) taken and used to create an almost identical or cloned email. The attachment or link within the email is replaced with a malicious version and then sent from an email address spoofed to appear to come from the original sender. It may claim to be a resend of the original or an updated version to the original. Typically this requires either the sender or recipient to have been previously hacked for the malicious third party to obtain the legitimate email.^[17][18]

Most methods of phishing use some form of technical deception designed to make a link in an email (and the spoofed website it leads to) appear to belong to the spoofed organization.^[19] Misspelled URLs or the use of subdomains are common tricks used by phishers. In the following example URL, http://www.yourbank.example.com/, it appears as though the URL will take you to the example section of the yourbank website; actually this URL points to the "yourbank" (i.e. phishing) section of the example website. Another common trick is to make the displayed text for a link (the text between the tags) suggest a reliable destination, when the link actually goes to the phishers' site. Many desktop email clients and web browsers will show a link's target URL in the status bar while hovering the mouse over it. This behavior, however, may in some circumstances be overridden by the phisher.[20] Equivalent mobile apps generally do not have this preview feature.^[21]

Internationalized domain names (IDN) can be exploited via IDN spoofing^[22] or homograph attacks,^[23] to create web addresses visually identical to a legitimate site, that lead instead to malicious version. Phishers have taken advantage of a similar risk, using open URL redirectors on the websites of trusted organizations to disguise malicious URLs with a trusted domain.^[24][25][26] Even digital certificates do not solve this problem because it is quite possible for a phisher to purchase a valid certificate and subsequently change content to spoof a genuine website, or, to host the phish site without SSL at all.^[27]

Phishers have sometimes used images instead of text to make it harder for anti-phishing filters to detect the text commonly used in phishing emails.^[28] In response, more sophisticated anti-phishing filters are able to recover hidden text in images using OCR (optical character recognition).^[29]

Some phishing scams use JavaScript commands in order to alter the address bar of the website they lead to.^[30] This is done either by placing a picture of a legitimate URL over the address bar, or by closing the original bar and opening up a new one with the legitimate URL.^[31]

An attacker can also potentially use flaws in a trusted website's own scripts against the victim.^[32] These types of attacks (known as cross-site scripting) are particularly problematic, because they direct the user to sign in at their bank or service's own web page, where everything from the web address to the security certificates appears correct. In reality, the link to the website is crafted to carry out the attack, making it very difficult to spot without specialist knowledge. Such a flaw was used in 2006 against PayPal.^[33]

To avoid anti-phishing techniques that scan websites for phishing-related text, phishers sometimes use Flash-based websites (a technique known as phlashing). These look much like the real website, but hide the text in a multimedia object.^[34]

Covert redirect is a subtle method to perform phishing attacks that makes links appear legitimate, but actually redirect a victim to an attacker's website. The flaw is usually masqueraded under a log-in popup based on an affected site's domain.^[35] It can affect OAuth 2.0 and OpenID based on well-known exploit parameters as well. This often makes use of open redirect and XSS vulnerabilities in the third-party application websites.^[36] Users may also be redirected to phishing websites covertly through malicious browser extensions.^[37]

Normal phishing attempts can be easy to spot because the malicious page's URL will usually be different from the real site link. For covert redirect, an attacker could use a real website instead by corrupting the site with a malicious login popup dialogue box. This makes covert redirect different from others.^[38][39]

For example, suppose a victim clicks a malicious phishing link beginning with Facebook. A popup window from Facebook will ask whether the victim would like to authorize the app. If the victim chooses to authorize the app, a "token" will be sent to the attacker and the victim's personal sensitive information could be exposed. These information may include the email address, birth date, contacts, and work history.^[36] In case the "token” has greater privilege, the attacker could obtain more sensitive information including the mailbox, online presence, and friends list. Worse still, the attacker may possibly control and operate the user’s account.^[40] Even if the victim does not choose to authorize the app, he or she will still get redirected to a website controlled by the attacker. This could potentially further compromise the victim.^[41]

This vulnerability was discovered by Wang Jing, a Mathematics Ph.D. student at School of Physical and Mathematical Sciences in Nanyang Technological University in Singapore.^[42] Covert redirect is a notable security flaw, though it is not a threat to the Internet worth significant attention.^[43]

Users can be encouraged to click on various kinds of unexpected content for a variety of technical and social reasons. For example, a malicious attachment might masquerade as a benign linked Google doc.^[44]

Alternatively users might be outraged by a fake news story, click a link and become infected.^[45]

Not all phishing attacks require a fake website. Messages that claimed to be from a bank told users to dial a phone number regarding problems with their bank accounts.^[46] Once the phone number (owned by the phisher, and provided by a voice over IP service) was dialed, prompts told users to enter their account numbers and PIN. Vishing (voice phishing) sometimes uses fake caller-ID data to give the appearance that calls come from a trusted organization.^[47]

A phishing technique was described in detail in a paper and presentation delivered to the 1987 International HP Users Group, Interex.^[49]

The term "phishing" is said to have been coined by the well known spammer and hacker in the mid-90s, Khan C Smith.^[50] The first recorded mention of the term is found in the hacking tool AOHell (according to its creator), which included a function for attempting to steal the passwords or financial details of America Online users.^[51][52]

Phishing on AOL was closely associated with the warez community that exchanged unlicensed software and the black hat hacking scene that perpetrated credit card fraud and other online crimes. AOL enforcement would detect words used in AOL chat rooms to suspend the accounts of individuals involved in counterfeiting software and trading stolen accounts. The term was used because "<><" is the single most common tag of HTML that was found in all chat transcripts naturally, and as such could not be detected or filtered by AOL staff. The symbol <>< was replaced for any wording that referred to stolen credit cards, accounts, or illegal activity. Since the symbol looked like a fish, and due to the popularity of phreaking it was adapted as "Phishing". AOHell, released in early 1995, was a program designed to hack AOL users by allowing the attacker to pose as an AOL staff member, and send an instant message to a potential victim, asking him to reveal his password.^[53] In order to lure the victim into giving up sensitive information, the message might include imperatives such as "verify your account" or "confirm billing information".

Once the victim had revealed the password, the attacker could access and use the victim's account for fraudulent purposes. Both phishing and warezing on AOL generally required custom-written programs, such as AOHell. Phishing became so prevalent on AOL that they added a line on all instant messages stating: "no one working at AOL will ask for your password or billing information". A user using both an AIM account and an AOL account from an ISP simultaneously could phish AOL members with relative impunity as internet AIM accounts could be used by non-AOL internet members and could not be actioned (i.e., reported to AOL TOS department for disciplinary action).^[54]. In late 1995, AOL crackers resorted to phishing for legitimate accounts after AOL brought in measures in late 1995 to prevent using fake, algorithmically generated credit card numbers to open accounts.^[55] Eventually, AOL's policy enforcement forced copyright infringement off AOL servers, and AOL promptly deactivate accounts involved in phishing, often before the victims could respond. The shutting down of the warez scene on AOL caused most phishers to leave the service.^[56]

"APWG Phishing Attack Trends Reports" ^[196] . Retrieved May 5, 2019.

There are anti-phishing websites which publish exact messages that have been recently circulating the internet, such as FraudWatch International and Millersmiles. Such sites often provide specific details about the particular messages.^[122][123]

As recently as 2007, the adoption of anti-phishing strategies by businesses needing to protect personal and financial information was low.^[124] Now there are several different techniques to combat phishing, including legislation and technology created specifically to protect against phishing. These techniques include steps that can be taken by individuals, as well as by organizations. Phone, web site, and email phishing can now be reported to authorities, as described below.

People can be trained to recognize phishing attempts, and to deal with them through a variety of approaches. Such education can be effective, especially where training emphasises conceptual knowledge^[125] and provides direct feedback.^[126][127]

Many organisations run regular simulated phishing campaigns targeting their staff to measure the effectiveness of their training.

People can take steps to avoid phishing attempts by slightly modifying their browsing habits.^[128] When contacted about an account needing to be "verified" (or any other topic used by phishers), it is a sensible precaution to contact the company from which the email apparently originates to check that the email is legitimate. Alternatively, the address that the individual knows is the company's genuine website can be typed into the address bar of the browser, rather than trusting any hyperlinks in the suspected phishing message.^[129]

Nearly all legitimate e-mail messages from companies to their customers contain an item of information that is not readily available to phishers. Some companies, for example PayPal, always address their customers by their username in emails, so if an email addresses the recipient in a generic fashion ("Dear PayPal customer") it is likely to be an attempt at phishing.^[130] Furthermore, PayPal offers various methods to determine spoof emails and advises users to forward suspicious emails to their spoof@PayPal.com domain to investigate and warn other customers. However it is it unsafe to assume that the presence of personal information alone guarantees that a message is legitimate,^[131] and some studies have shown that the presence of personal information does not significantly affect the success rate of phishing attacks;^[132] which suggests that most people do not pay attention to such details.

Emails from banks and credit card companies often include partial account numbers. However, recent research^[133] has shown that the public do not typically distinguish between the first few digits and the last few digits of an account number—a significant problem since the first few digits are often the same for all clients of a financial institution.

The Anti-Phishing Working Group produces regular report on trends in phishing attacks.^[134]

A wide range of technical approaches are available to prevent phishing attacks reaching users or to prevent them from successfully capturing sensitive information.

Specialized spam filters can reduce the number of phishing emails that reach their addressees' inboxes. These filters use a number of techniques including machine learning^[135] and natural language processing approaches to classify phishing emails,^[136][137] and reject email with forged addresses.^[3]

Another popular approach to fighting phishing is to maintain a list of known phishing sites and to check websites against the list. One such service is the Safe Browsing service.^[138] Web browsers such as Google Chrome, Internet Explorer 7, Mozilla Firefox 2.0, Safari 3.2, and Opera all contain this type of anti-phishing measure.^{[6][139][140][141][142]} Firefox 2 used Google anti-phishing software. Opera 9.1 uses live blacklists from Phishtank, cyscon and GeoTrust, as well as live whitelists from GeoTrust. Some implementations of this approach send the visited URLs to a central service to be checked, which has raised concerns about privacy.^[143] According to a report by Mozilla in late 2006, Firefox 2 was found to be more effective than Internet Explorer 7 at detecting fraudulent sites in a study by an independent software testing company.^[144]

An approach introduced in mid-2006 involves switching to a special DNS service that filters out known phishing domains: this will work with any browser,^[145] and is similar in principle to using a hosts file to block web adverts.

To mitigate the problem of phishing sites impersonating a victim site by embedding its images (such as logos), several site owners have altered the images to send a message to the visitor that a site may be fraudulent. The image may be moved to a new filename and the original permanently replaced, or a server can detect that the image was not requested as part of normal browsing, and instead send a warning image.^[146][147]

The Bank of America website^[148][149] is one of several that asks users to select a personal image (marketed as SiteKey) and displays this user-selected image with any forms that request a password. Users of the bank's online services are instructed to enter a password only when they see the image they selected. However, several studies suggest that few users refrain from entering their passwords when images are absent.^[150][151] In addition, this feature (like other forms of two-factor authentication) is susceptible to other attacks, such as those suffered by Scandinavian bank Nordea in late 2005,^[152] and Citibank in 2006.^[153]

A similar system, in which an automatically generated "Identity Cue" consisting of a colored word within a colored box is displayed to each website user, is in use at other financial institutions.^[154]

Security skins^[155][156] are a related technique that involves overlaying a user-selected image onto the login form as a visual cue that the form is legitimate. Unlike the website-based image schemes, however, the image itself is shared only between the user and the browser, and not between the user and the website. The scheme also relies on a mutual authentication protocol, which makes it less vulnerable to attacks that affect user-only authentication schemes.

Still another technique relies on a dynamic grid of images that is different for each login attempt. The user must identify the pictures that fit their pre-chosen categories (such as dogs, cars and flowers). Only after they have correctly identified the pictures that fit their categories are they allowed to enter their alphanumeric password to complete the login. Unlike the static images used on the Bank of America website, a dynamic image-based authentication method creates a one-time passcode for the login, requires active participation from the user, and is very difficult for a phishing website to correctly replicate because it would need to display a different grid of randomly generated images that includes the user's secret categories.^[157]

Several companies offer banks and other organizations likely to suffer from phishing scams round-the-clock services to monitor, analyze and assist in shutting down phishing websites.^[158] Individuals can contribute by reporting phishing to both volunteer and industry groups,^[159] such as cyscon or PhishTank.^[160] Individuals can also contribute by reporting phone phishing attempts to Phone Phishing, Federal Trade Commission.^[161] Phishing web pages and emails can be reported to Google.^[162][163] The Internet Crime Complaint Center noticeboard carries phishing and ransomware alerts.

Solutions have also emerged using the mobile phone^[164] (smartphone) as a second channel for verification and authorization of banking transactions.

Organisations can implement two factor or multi-factor authentication (MFA), which requires a user to use at least 2 factors when logging in. (For example, a user must both present a smart card and a password). This mitigates some risk, in the event of a successful phishing attack, the stolen password on its own cannot be reused to further breach the protected system. However, there are several attack methods which can defeat many of the typical systems.^[165] MFA schemes such as WebAuthn address this issue by design.

Organizations that prioritize security over convenience can require users of its computers to use an email client that redacts URLs from email messages, thus making it impossible for the reader of the email to click on a link, or even copy a URL. While this may result in an inconvenience, it does almost completely eliminate email phishing attacks.

An article in Forbes in August 2014 argues that the reason phishing problems persist even after a decade of anti-phishing technologies being sold is that phishing is "a technological medium to exploit human weaknesses" and that technology cannot fully compensate for human weaknesses.^[166]

On January 26, 2004, the U.S. Federal Trade Commission filed the first lawsuit against a suspected phisher. The defendant, a Californian teenager, allegedly created a webpage designed to look like the America Online website, and used it to steal credit card information.^[167] Other countries have followed this lead by tracing and arresting phishers. A phishing kingpin, Valdir Paulo de Almeida, was arrested in Brazil for leading one of the largest phishing crime rings, which in two years stole between US$18 million and US$37 million.^[168] UK authorities jailed two men in June 2005 for their role in a phishing scam,^[169] in a case connected to the U.S. Secret Service Operation Firewall, which targeted notorious "carder" websites.^[170] In 2006 eight people were arrested by Japanese police on suspicion of phishing fraud by creating bogus Yahoo Japan Web sites, netting themselves ¥100 million (US$870,000).^[171] The arrests continued in 2006 with the FBI Operation Cardkeeper detaining a gang of sixteen in the U.S. and Europe.^[172]

In the United States, Senator Patrick Leahy introduced the Anti-Phishing Act of 2005 in Congress on March 1, 2005. This bill, if it had been enacted into law, would have subjected criminals who created fake web sites and sent bogus emails in order to defraud consumers to fines of up to US$250,000 and prison terms of up to five years.^[173] The UK strengthened its legal arsenal against phishing with the Fraud Act 2006,^[174] which introduces a general offence of fraud that can carry up to a ten-year prison sentence, and prohibits the development or possession of phishing kits with intent to commit fraud.^[175]

Companies have also joined the effort to crack down on phishing. On March 31, 2005, Microsoft filed 117 federal lawsuits in the U.S. District Court for the Western District of Washington. The lawsuits accuse "John Doe" defendants of obtaining passwords and confidential information. March 2005 also saw a partnership between Microsoft and the Australian government teaching law enforcement officials how to combat various cyber crimes, including phishing.^[176] Microsoft announced a planned further 100 lawsuits outside the U.S. in March 2006,^[177] followed by the commencement, as of November 2006, of 129 lawsuits mixing criminal and civil actions.^[178] AOL reinforced its efforts against phishing^[179] in early 2006 with three lawsuits^[180] seeking a total of US$18 million under the 2005 amendments to the Virginia Computer Crimes Act,^[181][182] and Earthlink has joined in by helping to identify six men subsequently charged with phishing fraud in Connecticut.^[183]

In January 2007, Jeffrey Brett Goodin of California became the first defendant convicted by a jury under the provisions of the CAN-SPAM Act of 2003. He was found guilty of sending thousands of emails to America Online users, while posing as AOL's billing department, which prompted customers to submit personal and credit card information. Facing a possible 101 years in prison for the CAN-SPAM violation and ten other counts including wire fraud, the unauthorized use of credit cards, and the misuse of AOL's trademark, he was sentenced to serve 70 months. Goodin had been in custody since failing to appear for an earlier court hearing and began serving his prison term immediately.^{[184][185][186][187]}