The US case that will dictate the future of AI and data collection

The stakes are high because Kochava’s AI-assisted analysis and data acquisition practices, kept secret, are common in the global location data market. In addition to numerous lesser-known data brokers (intermediaries that collect data), the mobile data market includes larger players like Foursquare and data exchanges like Amazon’s AWS Data Exchange.

The FTC’s complaint against Kovacha, which was amended and declassified, shows that there is truth in what the company says: This can provide data for “any channel, any device, any audience”and buyers can “measure everything with Kochava.”

According to the FTC and class action lawsuits against Kochava on behalf of adults and children, the company secretly collects large amounts of data without notice or consent consumers’ personal This data is analyzed with AI, allowing Kovacha to predict and influence people’s behavior in diverse and invasive ways. Finally, this information is put up for sale. Faced with the accusations, Kochava has denied them all.

The problem of alleged mismanagement in the collection of sensitive data that is offered is not limited to Kovacha. This year, The FTC reached a settlement with data broker Outlogic, calling it the “first ban on the use and sale of location data sensitive.” The deal requires Outlogic to destroy the location data it stores on users and is prohibited from collecting or using such information to determine who enters and leaves sensitive locations, such as healthcare facilities, homeless shelters, and domestic abuse victims. , religious places, among others.
​
(Also: The ‘cold war’ over the chips that cars need: what are they used for?).

The federal entity maintains that Kochava sells a “360-degree profile” of people and advertises itself by promising that it has the ability to “connect “Precise geolocation data with email, demographics, devices, homes and channels.” In other words, Kochava takes location data, aggregates it with other data, and links it to consumer identities.

So the data that Kovacha sells reveals precise information about a person, such as their belonging to an ethnic group, sexual orientation, marital status, whether you have children, name and number of children, political tendency, level of education, economic situation, among others. But it also offers information on visits to hospitals, “reproductive health clinics, places of worship, shelters for homeless people and victims of domestic violence, addiction recovery centers.”

By selling such detailed data about people, the FTC believes that “Kochava is allowing third parties to identify individuals and expose them to threats.” of stigma, harassment, discrimination, loss of employment and even physical violence.”

As a lawyer who researches artificial intelligence and data privacy, these complaints show that US law is not up to date with the regulation of commercially available data and the governance of artificial intelligence.
​
​(Of interest: Google justifies the erroneous answers in AI Overview with ‘pointless searches’ and misuse of its AI).

According to the FTC, in addition to the data Kochava collects, the company does so in secret and then sells the ‘Kochava Collective’ information (precise geolocation data, complete individual consumption profiles, mobile app usage details, and “target segments”). audience”).

Kochava audience segments may be based on “behaviors” and sensitive information such as gender identity, political and religious affiliation, race, visits to hospitals and abortion clinics, and people’s medical information, such as menstruation and ovulation period, and even cancer treatments, says the federal entity.

When a Kovacha client selects certain audience segments, they can identify and target extremely specific groups. For example, people who do not identify as women or men and always choose “other” when asked about their gender identity, or all pregnant women who are African American and Muslim.

The FTC ensures that selected audience segments can be narrowed to a specific geographic region. And even, the specialization of data by georeference allows provide information on the inhabitants of a specific building.
​
Within the offered and selected segments, Kochava customers can obtain the name, home and email address, economic situation and job stability, and much more information. Who has a purchasing interest in the data? Organizations such as advertisers, insurers, and political campaigns that seek to classify and narrowly target people. The FTC also maintains that data can be purchased by people who want to harm others.

According to the FTC, Kochava acquires consumer data in two ways: through Kochava provides software devel
opment kits to application developers and directly from other data brokers. The FTC says those software development kits supplied by Kochava were used in more than 10,000 applications worldwide. These kits, which have Kochava encryption attached, collect large amounts of data and send it back to Kochava without the consumer being informed or consenting to the collection of the data.

(We recommend: Scientists launch a challenge for those who manage to communicate with animals using AI: the prize is 10 million dollars).

Another lawsuit against Kochava in California alleges similar charges of surreptitious data collection and analysis, and that Kochava sells customized data streams based on extremely sensitive and private information tailored to the needs of its customers.

The FTC complaint also illustrates how advanced AI tools are enabling a new phase in data analysis. Generative AI’s ability to process vast amounts of data is shaping what can be done and learned from mobile data in privacy-invading ways.. This includes inferring and revealing sensitive or legally protected information, such as medical records and images.

AI provides the ability to both know and predict almost anything about individuals and population groups, even very sensitive behaviors. It also allows behavior to be manipulated, inducing decisions supposedly in favor of specific users of the AI ​​tool.
​
This type of “AI coordinated manipulation” can supplant decision-making capabilities without people knowing.

In 2022, the FTC, which must ensure fair business practices and enforce the law, warned Kovacha that he was violating the law. Since then both sides have had some victories and losses in the ongoing case. Judge B. Lynn Winmill of the Idaho District Court, who is overseeing the case, dismissed the FTC’s first complaint and asked for more evidence. The commission filed an amended complaint that gave rise to many more specific allegations.
​
(Also read: WhatsApp will have new updates with artificial intelligence for June).

Winmill has not yet responded to a request from Kochava asking to dismiss the FTC complaints. At the beginning of this year, it became known that The parties are advancing their arguments. If a trial is reached, it is expected that it will be convened for 2025.
​
Probably for now, companies, privacy advocates and the political world are watching this case. Whatever ruling there is, combined with the proposed legislation and the FTC’s focus on generative AI, data and privacy, could mean big changes in how companies acquire data, the ways AI tools can be used to analyze data and what data can legally be used in machine- and human-based data analytics.

ANNE TOOMEY MCKENNA*
THE CONVERSATION**

Professor of Law, University of Richmond.
(**) It is a non-profit organization that seeks to share ideas and academic knowledge with the public. This article is reproduced here under license