November 30, 2023

Unlocking AI’s Power to Multiply Manufacturing Productivity

To capture AI’s promise, manufacturers must take steps to protect privacy and root out bias, particularly when they train their systems on data about employees.

Manufacturers have been conducting time and motion studies to increase efficiency and improve productivity for over a century. By doing so, they can often increase production speed and quality while improving working conditions.

Artificial intelligence (AI) systems could help manufacturers dramatically accelerate productivity gains, particularly because AI can enable them to analyze large datasets faster and with more precision and sophistication than ever before. The proliferation of digital technologies has enabled manufacturers to capture vast amounts of data about virtually every aspect of the production process. AI systems can analyze this data to identify bottlenecks and opportunities for improvement that would be difficult if not impossible to surface in any other way.

To unlock the power of AI, however, manufacturers will have to develop approaches that ensure they do not violate the privacy of employees or inadvertently promote bias or discrimination in their operations.

Privacy is a concern because a significant portion of the data that companies may capture and analyze is about employees, including location and biometric information. Bias is a concern because AI systems can inadvertently draw conclusions that could have discriminatory effects if operationalized.

Nevertheless, manufacturers that understand the challenge and take steps to ensure privacy and avoid discrimination can capture significant value from using AI to optimize productivity.

Location Data

Companies can gather data about the location and movements of employees using a variety of technologies, including personal or employer-issued mobile devices such as phones or wearables. These devices are typically configured with corporate software tailored to specific contexts. One company, for example, provides employees with wristbands that emit ultrasonic pulses, enabling it not only to track the location and movements of employees but to help employees locate inventory.

It is challenging for companies to strike a balance between the imperatives for productivity and protecting privacy, given the scarcity of case law on employer use of geolocation data. Despite the ambiguities, however, a few pivotal cases offer insights into where the lines may be drawn.

In the 2009 case of Haggins v. Verizon New England, Inc., a group of employees, represented by their union, sued their employer for requiring field technicians to carry company-issued cell phones so it could track their location. The employees claimed the policy violated their privacy rights under the Massachusetts Constitution and state law. The district court dismissed the case, and the First Circuit affirmed the dismissal. Notably, the district court introduced a precedent-setting balancing test, weighing employers’ interests in assessing employee effectiveness against employees’ rights to privacy.1

In 2015, an employee who was terminated for deleting tracking software from her mobile device pursued a wrongful termination suit. The case settled out of court, leaving no legal precedents but highlighting tensions between employee autonomy and employer surveillance.2

Some states have placed limitations on how and when employers may track the location of individuals. Notably, Section 637.7 of the California Penal Code restricts the use of electronic tracking devices unless the tracker is the registered owner or lessor of the vehicle.3

The significance of such state-level interventions is evident in a 2017 class action settlement, in which the US District Court for the Northern District of California ruled in favor of plaintiffs against the Bay Area Rapid Transit District for allegedly collecting location and other data from the personal devices of riders without clear user consent.4

In a watershed moment in 2018, the US Supreme Court ruled that collecting “cell site location information” without a warrant violated the Fourth Amendment, which protects against unreasonable searches and seizures.5 The court underscored the heightened privacy concerns associated with accessing personal location data compared to vehicle location data.

It is unclear whether tracking employees within the confines of private property, such as a factory, would render such surveillance less susceptible to scrutiny than  tracking employees out in the field. But as technology and legal frameworks evolve, the responsible use of location data in manufacturing facilities will require balancing operational efficiency with individual privacy rights.

Biometric Data

It is common for manufacturers to require employees to swipe a badge to identify themselves when entering or moving through a facility and to track hours on the job. Biometric data — gathered through fingerprints, palm prints, retinal scanners, or facial recognition — is increasingly used to manage access to sensitive areas and systems that warrant extra security.6

Biometric data can also be analyzed by AI systems as part of operational improvement efforts, but this can raise privacy concerns that can lead to legal challenges.

In 2019, for example, Walmart paid $10 million to settle a case alleging that it violated Illinois’ newly enacted privacy law. The plaintiffs in the case were required to scan their palms to unlock and lock their cash register drawer at the beginning and end of each shift. Walmart did not obtain written informed consent from the employees regarding the scan, and it failed to establish policies for the retention and destruction of biometric data as required by the Illinois Biometric Information Privacy Act.7  The court found the settlement to be fair, reasonable, and adequate.8

Bias and Discrimination

AI systems analyze data that may contain bias, and systems may sometimes come to conclusions that can inadvertently have discriminatory effects. AI systems that are trained on data gathered about a workforce that is imbalanced on certain factors such as gender, race, or age could come to false conclusions about the types of people who may perform best in certain roles.

In 2018, for example, Amazon had to shelve an automated resume review tool when it discovered that the tool preferred male candidates over female candidates.9 The bias allegedly emerged because the system had been trained on a dataset covering a 10-year period when the workforce was majority male. It can be challenging to root out this type of bias, but doing so is critical.

Three Steps for Manufacturers

The legal landscape is likely to change significantly over the next few years, especially as AI technology gets more sophisticated and the benefits of using it to analyze employee data increase. To limit their exposure to potential liability, manufacturers should consider taking the following steps:

  • Ensure compliance. Understand federal and state data privacy statutes, and design data collection and retention policies to ensure compliance. 
  • Secure consent. Develop clear and robust consent forms about data collection and require employees to sign them. The forms should be regularly updated and re-signed. 
  • Set parameters. Collect only data that is necessary to achieve the intended goals of the AI system that is being used. Less data is less risk.

 *     *     *

AI could bring tremendous productivity gains for manufacturers. But to capture its promise, they must understand and take steps to manage any legal or ethical challenges that may arise because of its use.


The authors thank Jessica Pauley, a 2023 summer associate, for her contribution to this article.


[1] Haggins v. Verizon New England, Inc., 648 F.3d 50 (1st Cir. 2011) (holding that under Massachusetts law, the reasonableness of an interference with an employee’s privacy is “a matter of custom and usage of the parties and their particular industry practices”).

[2] Arias v. Intermex Wire Transfer, 1:15-cv-01101 JLT (E.D. Cal. Sep. 14, 2015).

[3] Cal Penal Code § 637.7. Section 637.7 says that no person or entity “shall use an electronic tracking device to determine the location or movement of a person unless they [the person or entity doing the tracking] are the registered owner or lessor of the vehicle.” California Penal Code section 637.7 makes it a misdemeanor to install a GPS device to a person’s car without consent. While section 637.7 generally prohibits individuals from using electronic tracking devices to determine the location or movement of a person, an important exception allows registered owners, lessors, or lessees of vehicles to use electronic tracking devices to track their own vehicles. Connecticut, Delaware, and Texas also have laws requiring employee notice or consent before placing a GPS on an employee-owned car.

[4] Moreno v. S.F. Bay Area Rapid Transit Dist., 17-cv-02911-JSC, 2017 WL 6387764 (N.D. Cal. Dec. 14, 2017).

[5] Carpenter v. United States, 138 U.S. 2206 (2018) (holding a warrant is required to obtain cell-site location information in criminal cases).

[6] Te-Ping Chen, Workers Push Back as Companies Gather Fingerprints and Retina Scans, Wall Street Journal, March 27, 2019.

[7] Illinois Biometric Information Privacy Act, 740 ILCS 14/15(b)(3); 740 ILCS 14/15(b)(1)(b).

[8] Roach, et al. v. Walmart Inc., 2019-CH-01107, (Ill. Ct. Cl. Jan. 18, 2019).

[9] Jeffrey Dastin, Amazon scraps secret AI recruiting tool that showed bias against women, Reuters (Oct. 10, 2018).


This informational piece, which may be considered advertising under the ethical rules of certain jurisdictions, is provided on the understanding that it does not constitute the rendering of legal advice or other professional advice by Goodwin or its lawyers. Prior results do not guarantee a similar outcome.