Israel police warn tech companies to prepare for bots and AI election manipulation
With concerns growing over foreign interference and fake news, police are asking technology companies to identify coordinated networks, fake identities and automated political activity before they can distort the election.
The Israel Police is warning technology companies to prepare for the potential misuse of bots, fake accounts and automated systems during the upcoming Knesset election campaign, amid growing concerns over the spread of fake news and attempts to manipulate voters through digital platforms.
“Preparations to prevent the misuse of automation systems and bots during the national election campaign” is the title of a letter recently sent by Chief Boaz Blatt, head of the police’s Investigations and Intelligence Division, to technology companies. The division oversees the police’s cyber unit.
Blatt wrote that the police “intends to increase supervision and enforcement ahead of and during the election campaign, with the aim of ensuring the prevention and detection of offenses” relating to the “purity of the elections and the integrity of the democratic process.”
1 View gallery
PM Benjamin Netanyahu (right), Yashar Chairman Gadi Eisenkot
(Shalev Shalom, Shaul Golan)
Blatt stressed that his approach to the companies did not stem from any suspicion against them. Rather, he said the purpose was to raise awareness, establish channels of cooperation with the police and encourage companies to prepare in advance to reduce risks during the election period.
“The Israel Police sees technology companies as key partners in protecting the digital space and public trust, and encourages responsible and supervised conduct during the election period,” he wrote.
The police chief warned of the potential for technology systems to be used illegally “both for the dissemination of false content and for illegal distribution methods that create false representations regarding the source or veracity of the information.”
The letter provides technology companies with a list of warning signs they should watch for, along with recommendations intended to help them avoid becoming unwitting tools for criminal activity involving automation systems, bots and networks of fake accounts.
According to Blatt, such systems can enable the rapid and coordinated distribution of content on a large scale, including the dissemination of false information, the creation of artificial support for political messages, impersonation and the concealment of the identities of those operating the networks.
He also warned that “assistance in concealing identity, circumventing restrictions, or continuing to provide service after a real suspicion has arisen may establish legal liability for the company, depending on the circumstances.” Avoiding the clarification of a genuine suspicion could also be considered when assessing liability, he added, while emphasizing that each case must be examined individually and in accordance with the law.
The warnings come as Israeli authorities are increasingly concerned about the possibility that networks of fake users and bots could be deployed for political promotion or to spread false information during the election campaign. This time, however, those capabilities are being combined with a new and potentially more powerful tool: artificial intelligence.
Concerns over possible foreign interference, including by Iran, have added another dimension to the preparations. Israeli political actors are also expected to use increasingly sophisticated digital tools during the campaign.
The police and Shin Bet are preparing to monitor attempts by hostile entities and foreign countries to conduct influence operations aimed at affecting the Knesset elections.
At a discussion held at the end of July by a subcommittee of the Knesset Foreign Affairs and Defense Committee, Supreme Court Justice Noam Sohlberg, chairman of the Central Elections Committee, and Shin Bet chief David Zini participated in discussions about election security. Zini warned of attempts by foreign countries to influence the elections, particularly through cyberspace and the use of bots on social networks.
Against that backdrop, the police are also working directly with technology companies, sending them letters and planning meetings with companies to discuss the issue.
“The election system is expected to entail technological and other challenges that we have not faced in the past,” Blatt wrote. He warned that criminal offenses could undermine “the integrity of the election procedures, the purity of the elections and, by the way, the public’s trust in the democratic process,” adding that the police were preparing to deal with those threats.
Four categories of warning signs
The police letter divides the warning signs that technology companies should monitor into four broad categories.
The first involves communications or usage patterns characterized by the coordinated operation of large numbers of accounts, identities or telephone numbers that falsely represent who or what they are. Other warning signs include the creation of content, responses or inquiries that appear genuine but are generated on an unusual scale, as well as political messages presented as independent human activity even though they originate from an automated and coordinated system.
The second category involves the use of false identities or information that does not match the identity of the customer. Particular attention should be paid when there are attempts to conceal the identity of the customer, the person financing an activity or its ultimate beneficiary, including through intermediaries, shell companies, foreign entities or chains of suppliers.
The third concerns suspicious requests involving personal databases, telephone-number lists or user data whose origins are unclear.
The police also singled out impersonation of individuals, political candidates, parties, public authorities or media organizations. Another concern is the dissemination of false or misleading information about basic election procedures, including the location of polling stations, the date of the election, voting methods or the identity of people participating in the electoral process.
The fourth category concerns attempts to circumvent the safeguards built into technology platforms. These include efforts to bypass rate limits, verification mechanisms, security measures and other tools designed to identify inauthentic activity.
Sharp and unexplained changes in the volume of activity, the number of accounts, geographic areas of activity, the content being distributed or the identity of the entities managing the engagement should also raise questions, according to the police.
Blatt emphasized that no single warning sign necessarily proves that prohibited activity is taking place.
“However, an accumulation of warning signs, contradictions in the client's explanations, avoidance of providing basic details, or continuation of activity after a real concern has arisen regarding its purpose, require increased scrutiny,” he wrote.
In appropriate cases, he said, companies should consider restricting, suspending or terminating services and reporting suspicious activity in accordance with the law and their contractual terms.
Where there are concerns about interference with the integrity of the elections or the commission of computer-related crimes, Blatt called on companies to contact the police.
The approach reflects a broader challenge facing Israeli authorities ahead of the election: the infrastructure used to communicate with voters is increasingly controlled not by governments or political parties, but by technology platforms and companies whose systems can be used to create, amplify and distribute content at a scale that would have been difficult to imagine in previous election campaigns.
The police are now effectively asking those companies to become another line of defense for the electoral system, before an election in which the distinction between genuine human activity and machine-generated influence could become increasingly difficult to establish.
