¶ 1. The John K. Maclver Institute for Public Policy, Inc. and Brian Fraley, hereinafter collectively "the Institute," appeal from a circuit court order denying their request for a writ of mandamus directing Wisconsin State Senator Jon Erpenbach to disclose certain public policy related e-mails sent to him — without redaction of information identifying the sender or the e-mail address. The Institute contends the e-mails it seeks must be released without redaction of the identifying information because they are public records and the public interest in redacting the information is not greater than the public interest in disclosing it. Erpenbach responds that his decision to keep the identifying information confidential is in compliance with custom and practice of the Wisconsin Senate and, therefore, even if the decision is inconsistent with the open records law, the matter before us is nonjusticiable. He also contends the information sought by the Institute is "purely personal" and therefore not subject to disclosure under Schill v. Wisconsin Rapids School District, 2010 WI 86, 327 Wis. 2d 572, 786 N.W.2d 177. He further asserts that the public interest in nondisclosure of the information outweighs the public interest in disclosure. We conclude that this matter is justiciable, the redacted information is not "purely personal," and the public interest in keeping the identifying information secret does not outweigh the public interest in disclosure. We reverse.
BACKGROUND
¶ 2. In February 2011, legislation was introduced in the Wisconsin Legislature proposing substantial changes to Wisconsin's collective bargaining laws. The *66Assembly and Senate passed the legislation, which eventually became 2011 Wis. Act 10, and the governor signed it into law in mid-March 2011. While the legislation was under consideration, and for some time thereafter, passions ran high among various interested parties and individuals. Protests and rallies, as well as threats against and recall elections of public officials, flowed from these passions.
¶ 3. Upon that stage, the following relevant facts are undisputed. On March 24, 2011, the Institute submitted a written request to Erpenbach seeking copies of all correspondence to and from him related to the collective bargaining changes. On April 18, 2011, Erpenbach informed the Institute that some of the requested documents were available to be picked up, but that he had redacted personal contact information or personally identifiable information, including last names and e-mail addresses, in certain of the documents.
¶ 4. In the months that followed, the Institute narrowed its request to only seeking unredacted e-mails sent from state and local government e-mail accounts, sending Erpenbach its final written request for such documents on November 2, 2011. On November 13, 2011, Erpenbach responded that he would not provide the "public e-mail addresses of state employees and other public employees."
¶ 5. Through his April and November responses, Erpenbach expressed that he was refusing to provide the redacted information because the communications were "purely personal" under Schill. He also asserted that the public interest in nondisclosure of the information outweighed the public interest in disclosure because nondisclosure protects against the "potential for threats, harassment and reprisals" against e-mail *67senders, respects senders' privacy and rights to free speech and to petition the government, and guards against the potential "chilling effect" of disclosure on future citizen communications.
¶ 6. The Institute filed this lawsuit seeking a writ of mandamus to compel Erpenbach to allow inspection and copying of the requested correspondence without redaction of identifying information. The circuit court denied the writ request, concluding that "[w]hile this court may not have arrived at the same conclusion as did Senator Erpenbach, it is required by case law to accord deference to his judgment." The Institute appeals. Additional facts are set forth as necessary.
DISCUSSION
Justiciability
¶ 7. Asa threshold matter, we address Erpenbach's contention that this mandamus action is not properly before the courts. Erpenbach asserts that it has long been a custom and practice of the Wisconsin Senate "to leave it up to each individual Senator whether to disclose personally identifiable information regarding constituents who contact the Senator."1 As a result, he contends that such disclosure decisions are a matter of the Senate's "Rules of Proceeding," and that pursuant to article iy section 8 of the Wisconsin Constitution, the courts "may not question the wisdom, or pass on the *68validity, of the rule of proceeding," even if a senator's nondisclosure decision is inconsistent with the open records law. We conclude that this matter is justiciable because Erpenbach's nondisclosure decision does not implicate the Senate's Rules of Proceeding.
¶ 8. Article IY section 8 of the Wisconsin Constitution provides: "Each house may determine the rules of its own proceedings." Our supreme court has defined such rules as those having "to do with the process the legislature uses to propose or pass legislation." Custodian of Records for Legislative Tech. Servs. Bureau v. State, 2004 WI 65, ¶¶ 29-30, 272 Wis. 2d 208, 680 N.W.2d 792. Courts will not "intermeddle" in "purely internal legislative proceedings." Milwaukee Journal Sentinel v. DOA, 2009 WI 79, ¶ 18, 319 Wis. 2d 439, 768 N.W.2d 700. Whether Erpenbach's nondisclosure decision implicates article I\[ section 8, is a constitutional question we review de novo. Custodian of Records, 272 Wis. 2d 208, ¶¶ 6, 29.
¶ 9. We are guided by our supreme court's decision in Custodian of Records. In that John Doe proceeding, the director of the Wisconsin Legislature's Legislative Technology Service Bureau (LTSB) sought to quash a judge's subpoena seeking electronically stored communications on all of the LTSB servers, or, in the alternative, "all 'documents' for certain named legislators [and] their aides." Id., ¶¶ 2, 4, 5. Among other items maintained on the servers were "legislators', constituents' and service agency e-mails." Id., % 3. The LTSB director contended that Wis. Stat. § 13.96 (2001-02),2 which required the LTSB to "at all times *69observe the confidential nature" of the data, was an article section 8 rule of proceeding that only the legislature was permitted to interpret and that the question of the enforceability of the subpoena was a nonjusticiable political issue. Custodian of Records, 272 Wis. 2d 208, ¶¶ 13, 24, 27. The court concluded that § 13.96 (2001-02) was not a rule of proceeding, noting that the provision "simply provides for assistance with electronic data and for an electronic storage closet for communications created or received by legislators and other employees of the legislature" and has "nothing to do with the process the legislature uses to propose or pass legislation." Custodian of Records, 272 Wis. 2d 208, ¶¶ 29-30.
¶ 10. We observe that the legislature wrote the open records law to apply to "elected official [s]" generally, without any special exception for individual state legislators or houses of the legislature. See Wis. Stat. § 19.32(1). Erpenbach has identified no difference in how our laws treat policy-related correspondence to legislators and how it treats similar correspondence to any other elected state officials. We presume that, like members of the legislature, the governor and perhaps even the attorney general, both elected officials, also received correspondence encouraging them to act in favor of or in opposition to the public policy objectives of Act 10.3 And, while consideration of public opinion regarding policy matters is essential to any thoughtful legislator, in his or her role as an individual "authority" under the open records law, see 19.32(1), the legislator maintains custody of correspondence like other elected officials and, like the LTSB in Custodian of Records, *70acts to a certain extent as "an electronic storage closet for communications" received by the legislator. See Custodian of Records, 272 Wis. 2d 208, ¶ 29.
¶ 11. Simply put, the Institute's request for a writ does not relate to "purely internal legislative proceedings" or implicate the methods or "process the legislature uses to propose or pass legislation." Cf. State ex rel. La Follette v. Stitt, 114 Wis. 2d 358, 364, 338 N.W.2d 684 (1983) (holding that the courts "will not determine whether internal operating rules or procedural statutes have been complied with by the legislature in the course of its enactments" (emphasis added)). The matter before us is justiciable.
Disclosure
¶ 12. After an in camera review of the contested e-mails, the circuit court concluded that "[w]hile this court may not have arrived at the same conclusion [regarding the balancing test] as did Senator Erpenbach, it is required by case law to accord deference to his judgment." We disagree with the circuit court's view of the law.
¶ 13. When addressing an open records request, a records custodian must make the initial decisions on whether a requested item is a "record" and whether any statutory or common law exceptions to disclosure apply. See Zellner v. Cedarburg Sch. Dist., 2007 WI 53, ¶¶ 23-31, 300 Wis. 2d 290, 731 N.W.2d 240; Linzmeyer v. Forcey, 2002 WI 84, ¶ 11, 254 Wis. 2d 306, 646 N.W.2d 811. If the custodian determines that the item is a record and no exceptions apply, the custodian must then conduct a balancing test to "weigh the competing *71interests involved and determine whether permitting inspection would result in harm to the public interest which outweighs the legislative policy recognizing the public interest in allowing inspection." Osborn v. Board of Regents of the Univ. of Wis. Sys., 2002 WI 83, ¶ 15, 254 Wis. 2d 266, 647 N.W.2d 158 (citation omitted).
¶ 14. If the custodian's decision is challenged, however, a court must make its own independent decisions regarding these matters, including the balancing test. Hempel v. City of Baraboo, 2005 WI 120, ¶ 21, 284 Wis. 2d 162, 699 N.W.2d 551; Osborn, 254 Wis. 2d 266, ¶ 12; Seifert v. School Dist. of Sheboygan Falls, 2007 WI App 207, ¶ 16, 305 Wis. 2d 582, 740 N.W.2d 177; Milwaukee Journal v. Call, 153 Wis. 2d 313, 317, 450 N.W.2d 515 (Ct. App. 1989). "The duty of the custodian is to specify reasons for nondisclosure and the court's role is to decide whether the reasons asserted are sufficient." Fox v. Bock, 149 Wis. 2d 403, 416, 438 N.W.2d 589 (1989). If the custodian states no reason or insufficient reasons for refusing to disclose the information, the writ of mandamus compelling disclosure must issue. Osborn, 254 Wis. 2d 266, ¶ 16. A court should apply the balancing test "when the record custodian has refused to produce the record, in order to evaluate the merits of the custodian's decision." Milwaukee Journal Sentinel, 319 Wis. 2d 439, ¶ 55. Where, as here, the relevant facts are undisputed, we review de novo a custodian's balancing decision of whether the public interest in nondisclosure of the challenged information outweighs the public interest in disclosure. See id., ¶ 14; Linzmeyer, 254 Wis. 2d 306, ¶ 12 (citing Woznicki v. Erickson, 202 Wis. 2d 178, 549 N.W.2d 699 (1996)). It is the burden of the party seeking nondisclosure to show that "public interests favoring *72secrecy outweigh those favoring disclosure." C.L. v. Edson, 140 Wis. 2d 168, 182, 409 N.W.2d 417 (Ct. App. 1987); see also Fox, 149 Wis. 2d at 416. Access is only to be denied "in an exceptional case." Wis. Stat. § 19.31.
¶ 15. Erpenbach suggests we should give a "heightened level of deference" to his decision as custodian to redact the challenged information because he is an elected lawmaker and the environment in which he made the nondisclosure decision was one of "unprecedented circumstances." While we recognize that we must "examin[e] ... all the relevant factors, considered in the context of the particular circumstances," Seifert, 305 Wis. 2d 582, ¶ 31, Erpenbach provides no law supporting his suggestion that legislator-custodians should be afforded deference in their disclosure decisions that is not afforded to nonlegislator-custodians. We will not take it upon ourselves to create a rule treating legislators differently from other elected or nonelected records custodians. Our review of Erpenbach's balancing decision remains de novo.4
¶ 16. The legislature has declared that:
[I]t is . .. the public policy of this state that all persons are entitled to the greatest possible information regarding the affairs of government and the official acts of those officers and employees who represent them. ...
*73[P]roviding persons with such information is declared to be an essential function of a representative government .... To that end, [Wis. Stat. §§ 19.32 to 19.37 shall be construed in every instance with a presumption of complete public access, consistent with the conduct of governmental business. The denial of public access generally is contrary to the public interest, and only in an exceptional case may access be denied.
Wis. Stat. § 19.31 (emphasis added). This statement "is one of the strongest declarations of policy to be found in the Wisconsin statutes." See Zellner, 300 Wis. 2d 290, ¶ 49. Following this declaration, Wisconsin maintains a "strong presumption of complete openness with regard to public records." Id., ¶ 55.
¶ 17. Here, the Institute requested and continues to seek from Erpenbach unredacted e-mails sent from state and local government e-mail accounts. The Institute contends Erpenbach's refusal to produce the e-mails without redaction of information identifying the senders and the e-mail addresses from where they were sent is in violation of the open records law because the e-mails are "records" under the law, the law creates a broad presumption that such records are to be disclosed, and under the balancing test, the public interest in nondisclosure of the redacted information does not outweigh the public interest in disclosure. Erpenbach argues that under Schill the redacted information is not subject to disclosure because it is "purely personal," and he further contends the public interest in nondisclosure outweighs the public interest in disclosure. We agree with the Institute.
¶ 18. Erpenbach does not dispute that the e-mails themselves are public records, nor could he successfully do so. The e-mails were sent to an elected lawmaker *74generally for the purpose of influencing the lawmaker's position on public policy, and they were maintained on a government e-mail system. Erpenbach acknowledges in his appellate brief that the e-mails "concernQ pending legislation that would impact every tax paying citizen of the State" and that the "concerns of citizens helped shape his official acts" concerning collective bargaining legislation. Without question, the e-mails relate to "the affairs of government," "the official acts" of a public officer, and have a clear connection to a government function — enactment of public policy. See Wis. Stat. § 19.31; Schill, 327 Wis. 2d 572, ¶¶ 22, 80-81, 119 (plurality opinion). The e-mails are public records.5
¶ 19. Erpenbach asserts, however, that the redacted portions of the e-mails are "purely personal" and therefore not subject to disclosure. Personal finance or health information of e-mail senders is not at issue,6 but only information such as names and e-mail ad*75dresses, which would identify the senders and from where the e-mails were sent. Erpenbach has identified no statutory or common law exception to disclosure of this identifying information, but contends the information has "no connection" to his ' "official acts'. .. [or] to his 'government function,'" and, therefore, is not subject to disclosure. The Institute argues that "who" is attempting to influence a legislator is an "integral part" of the communication itself and relates to the affairs of government and Erpenbach's official acts, and it further asserts that from "where" the e-mails were sent is also of interest to the public. Again, we agree with the Institute.
¶ 20. Public awareness of "who" is attempting to influence public policy is essential for effective oversight of our government. For example, if a person or group of persons who has played a significant role in an elected official's election — by way of campaign contributions or other support — contacts a lawmaker in favor of or opposed to proposed legislation, knowledge of that information is in the public interest; perhaps even more so if the person or group also stands to benefit from or is at risk of being harmed by the legislation. Disclosure of information identifying the sender may assist in revealing such a connection. Here, for example, the circuit court observed that "Act 10 personally affected all government employees" and Erpenbach acknowledged in his affidavit that Act 10 "directly affected [the] rights and obligations" of the public employees who e-mailed him. The Institute asserts that "there are a number of form e-mails that make up a substantial amount of the e-mails received by Senator Erpenbach" *76and that this, along with the content of the e-mails, suggests coordinated activity by public employee unions to affect the outcome of Act 10. Disclosure of the redacted information sought here can provide the public with knowledge and insight regarding who was attempting, either individually or in an organized fashion, to influence the public policy changes under consideration and thereby assist the public in performing its important government oversight function. See, e.g., Milwaukee Journal Sentinel v. City of Milwaukee, 2012 WI 65, ¶ 4, 341 Wis. 2d 607, 815 N.W.2d 367 (the open records law "reaffirms that the people have not only the opportunity but also the right to know what the government is doing and to monitor the government"); Schill, 327 Wis. 2d 572, ¶ 2 (plurality opinion) ("The right of the people to monitor the people's business is one of the core principles of democracy." (quoting Editorial, Shine Light on Public Records, Wisconsin State Journal, Mar. 14, 2010, at B1)). Whether government employees, another public official, a lobbyist, the CEO or employees of a corporation, the president or members of a union, or other individuals supporting or opposing a particular interest, awareness of who is attempting to influence public policy is of significant interest to the public.
¶ 21. It is also of public interest to know from "where" the sender is attempting to influence public policy. Whether a communication is sent to a public official from a source that appears associated with a particular unit of government (such as Milwaukee County or Waukesha School District), a private entity (such as Northwestern Mutual Life or Marquette University), or a nonprofit organization (such as American Red Cross or Clean Wisconsin, Inc.), or from individuals who may be associated with a specific interest or *77particular area of the state, from "where" a communication is sent further assists the public in understanding who is attempting to influence public policy and why. Thus, the redacted information identifying "who" sent e-mails attempting to influence public policy and from "where" the e-mails were sent is not "purely personal," and the public has a strong interest in disclosure of such information.
¶ 22. Having discussed the public interest in disclosure of the redacted information, we turn to the second part of the balancing test — the public interest in keeping the information secret. Erpenbach justifies his redactions on the grounds that they protect the e-mail senders against "unwanted harassment, [and] threat of reprisals," respect senders' privacy and rights to free speech and to petition the government, and guard against the "chilling effect" of disclosure on future citizen communications. Because Erpenbach develops no free speech or petition rights arguments separate from his contentions regarding his other balancing test considerations, we do not address those factors in any manner distinct from the other considerations.
¶ 23. Erpenbach places much emphasis on the "nuclear environment" which existed in and around the state Capitol building while collective bargaining changes were being considered and even after their enactment, and, from this, argues that senders of the e-mails generally could face threats, harassment or reprisals in one form or another. While Erpenbach has identified threats and harassment levied against public officials and police officers at the Capitol building itself around the time Act 10 was under consideration, he has identified no instances of actual threats, harassment or reprisals against concerned citizens away from the Capitol building who merely communicated their posi*78tion regarding the public policy changes via correspondence. And while he asserts generally in his affidavit that many of the communications he received expressed concern about reprisal, he has failed to identify any e-mails in the record supporting that assertion, and our own review of the e-mails has not revealed any such concerns by the senders. In short, Erpenbach has demonstrated no reasonable probability that citizens merely corresponding with lawmakers regarding their position on Act 10 would be subjected to negative repercussions for sharing their views regarding the legislation.
¶ 24. Public policy changes which individuals and groups feel passionately about have been enacted before and will be enacted again in the future. Erpenbach's generalized concern of possible threats, harassment or reprisals could apply equally to any controversial public policy. Apparently recognizing that there is always some risk to a citizen who voices an opinion on an issue about which he or she cares, the Institute correctly observes that "many political disputes are contentious (abortion, the death penalty, school choice, right to work, same sex marriage, etc.)." The Institute argues that adopting Erpenbach's position would create a rule which would result in the "avoid[ance] [of] disclosure on all issues that the public cares most about: the contentious ones." We agree, and decline to adopt such a rule, which would be contrary to the public interest and the presumption of openness.
¶ 25. We gain guidance from the United States Supreme Court's recent decision in Doe v. Reed, 561 U.S. 186, 130 S. Ct. 2811 (2010), cited by both parties. Similar to signers of a referendum petition in Reed, the senders of the e-mails at issue in this case were attempting to affect state public policy. In Reed, thou*79sands of citizens signed a petition to place on the ballot a statewide referendum which could overturn a new law signed by the governor extending certain benefits to same-sex couples. Id., 130 S. Ct. at 2816. Pursuant to that state's public records law, persons sought copies of the petition, which included the names and home addresses of signers. Id. In addition, two entities issued a press release stating their plans to post the names of the signers online in a searchable format. Id. The petition sponsor and certain signers filed a lawsuit seeking to enjoin the release of the petition. Id. Plaintiffs asserted that the public records law was unconstitutional as applied to the petition because there was a reasonable probability signers would be "subjected to threats, harassment, and reprisals," due to the purported plan to identify signers on the Internet and additional plans by groups to encourage other citizens to seek out petition signers. Id. at 2816, 2820. While the question the Reed Court ultimately addressed was whether disclosure of the referendum petitions in general would violate the First Amendment, id. at 2815, the Court explained that those resisting disclosure can prevail "if they can show 'a reasonable probability that the compelled disclosure [of personal information] will subject them to threats, harassment, or reprisals from either Government officials or private parties," id. at 2821 (quoting Buckley v. Valeo, 424 U.S. 1, 74 (1976)); see also Citizens United v. FEC, 130 S. Ct. 876, 915 (2010). Similarly, in Lassa v. Rongstad, 2006 WI 105, 294 Wis. 2d 187, 718 N.W.2d 673, our supreme court denied, in the First Amendment context, a litigant's request to keep information identifying donors to his organization a secret, noting that "[a]lthough a record of past harassment or reprisals is not always necessary, *80[the litigant] has not otherwise established a reasonable probability of such chilling effects." Id., ¶¶ 67-70.
¶ 26. While Erpenbach correctly asserts that the possibility of threats, harassment or reprisals alone is a legitimate consideration for a custodian, the public interest weight given to such a consideration increases or decreases depending upon the likelihood of threats, harassment or reprisals actually occurring. While there certainly is some possibility that citizens away from the Capitol building who sent e-mails could have faced threats, harassment or reprisals when Erpenbach made his decisions against disclosure,7 we cannot accord significant weight to this consideration due to his failure to establish a reasonable probability of any such harm. See Hempel, 284 Wis. 2d 162, ¶ 79 ("Factual support for the custodian's reasoning is likely to strengthen the custodian's case before a circuit court.").
¶ 27. Erpenbach also contends his decision to redact identifying information on the e-mails respects the senders' privacy and prevents a "chilling effect" that would otherwise occur on future communications if citizens became aware their names and comments will become a public record. We are unpersuaded.
¶ 28. As the Institute points out, the senders of the e-mails "chose ... to speak and petition in a manner that made their identity plain and easily-identifiable by including their names within their messages and using a method of communication that transmits the identifying account of the sender." In this day and age, it would be unreasonable for a person sending an e-mail to a lawmaker in an attempt to influence public policy *81to believe that the e-mail and all the information therein might not be seen by persons other than just the lawmaker.
¶ 29. Further, Erpenbach has not directed us to any portion of the record demonstrating that e-mail senders had an expectation their e-mails, or portions thereof, would be kept secret, and our own review of the e-mails has not revealed any such expectation.8 Indeed, Erpenbach has identified no law suggesting he himself was or is precluded from sharing these e-mails, in unredacted form, with whomever he chooses. As the circuit court concluded, "senders must realize that the recipient of an e-mail may print, forward, or otherwise disclose the contents of the communication, unless otherwise privileged." The circuit court observed that the term "private" is "oxymoronic with sending an e-mail to a public official concerning a public matter." We agree with the circuit court's conclusion that the senders of the e-mails in this case "had no expectation that the e-mails would remain private," and certainly not a reasonable expectation.
¶ 30. Erpenbach argues that disclosure of the e-mail senders' identifying information will "chill" citizens from communicating with legislators. As support, he cites to an affidavit of his expert, a University of Wisconsin political science professor, wherein the professor opines that disclosing citizens' personally identifiable information together with the content of their communication to a lawmaker will deter such communication. The professor's conclusions are based in substantial part upon his review of a study related to the
*82Iowa Caucuses showing that some persons were deterred from attending a caucus to vote for a candidate when they were informed that their vote would be public. The professor's opinion would be of more import if the issue before us was simply citizens communicating versus not communicating their views to a legislator, like attending or not attending a caucus to vote for a candidate. The issue before us, however, revolves around the medium of communication utilized; and nothing about our holding prevents citizens from sharing their views with public officials via phone or in-person communication, two routinely utilized methods which allow citizens to share their views with a public official without necessarily creating a public record related to those views.
¶ 31. It is without question laudable and desirable for citizens to contact elected officials to express their opinions regarding public policy. There is limited reason for us to conclude, however, that release of the e-mails at issue without redaction of identifying information will chill persons from ultimately communicating their views to a public official. If a citizen has a genuine concern about his or her views becoming public, he or she need not express such views through means which create a public record.
¶ 32. Transparency and oversight are essential to honest, ethical governance. Erpenbach has not met his burden of establishing that the public interest in nondisclosure of the redacted information outweighs the significant public interest in disclosure. Accordingly, he has not overcome the "strong presumption of complete openness" with regard to the e-mails. Zellner, 300 Wis. 2d 290, ¶ 55.
*83 CONCLUSION
¶ 33. For the foregoing reasons, we reverse and remand with directions to the circuit court to order Erpenbach to release the requested records without redaction of identifying information such as the name and e-mail address of the sender. Additionally, on remand, we direct the circuit court to determine the appropriate costs and fees to be awarded the Institute pursuant to Wis. Stat. § 19.37(2)(a).
By the Court. — Order reversed and cause remanded with directions.
Erpenbach references "constituents." By this we assume he means individuals who reside in his state senate district. Erpenbach has made no showing that he limited his redactions to e-mails sent only by his constituents, nor does it appear from the record that he could do so. Some of the e-mails at issue include home addresses of the e-mail sender and show e-mails from nonconstituents as well as constituents.
All references to the Wisconsin Statutes are to the 2011-12 version unless otherwise noted.
In fact, our review of the contested e-mails revealed that some of the e-mails sent to Erpenbach were also sent to the governor.
Besides being the law, de novo review makes sense because it is unlikely that in all cases records custodians will be as neutral as the courts. Records requested well may relate to the custodian himself or herself, as in this case, or to the custodian's boss, or a coworker the custodian engages with on a regular basis. In short, a custodian personally may view a records request as being favorable or unfavorable to his or her own interests or those of someone close to him or her. The courts generally provide a more disinterested forum.
Wisconsin Stat. § 19.32(2) provides that "records" include "any material on which ... electromagnetic information is recorded or preserved, regardless of physical form or characteristics, which has been created or is being kept by an authority." See also Schill v. Wisconsin Rapids Sch. Dist., 2010 WI86, ¶ 56 (plurality opinion of Justices Abrahamson, Crooks and Prosser), ¶¶ 149-52 (Bradley, J., concurring), ¶¶ 173-75 (Gableman J., concurring), ¶ 206 (Roggensack, J., dissenting) (all discussing e-mails as records), 327 Wis. 2d 572, 786 N.W.2d 177. Section 19.32(1) defines "authority" to include an "elected official" who has custody of a record.
We recognize that within a communication attempting to influence public policy there may be information which could be properly redacted, either because it is purely personal or because the public interest in nondisclosure outweighs the public interest in disclosure. For example, if a citizen e-mails an elected official in opposition to or in favor of abortion-related laws and therein references an abortion she had when she was *75fifteen years old, it would be reasonable for a custodian to conclude that such a limited, very personal reference could be redacted.
We note that Erpenbach's final written rejection (prior to this lawsuit) of the Institute's request for unredacted records came eight months after enactment of Act 10.
We further observe that, in many cases, e-mail senders did not send their e-mails just to Erpenbach, but also sent the same e-mails to other legislators, the governor, and other individuals, including "undisclosed recipients."