In the first part of this article, Marion Heathcote, Chair of the MARQUES IP Emerging Issues Team, noted that the Team asked: if AI understands us better than we understand ourselves, why not let it handle our purchasing decisions entirely? What would this mean for trade mark law?

To answer those questions, we are reproducing the Team’s original 2022 draft article, penned by Michal Hasiow, on ‘Navigating the pitfalls of brand control in product and packaging personalisation’.

The previously published part 1 introduced the concept. In this second part, we look at how brands can retain (or regain) control.

Boundaries and safe words

While brands’ trust vis-a-vis their customers can lead to a richer customer experience, it should have its limits. Personalisation options can be kept in check by offering a curated catalogue of choices – allowing consumers to select from a predefined set of features. This approach, focusing on design elements such as colours, shapes or ribbons, significantly reduces risks. However, some combinations (such as national flag colours) may still spark controversy, depending on the context.

Allowing predefined texts, such as initials or short phrases, is generally low risk. But while these options seem safe, they still leave room for problematic combinations, such as “SS” or “KGB”. The real challenge arises with freely generated content, such as longer text or photos, which are harder to control and carry higher risks.

Automated filtering systems help screen inappropriate terms, but they come with technical limitations. Too much filtering can stifle creativity, while too little allows harmful content to slip through. Finding the right balance is essential to mitigate the risk of brand damage.

For instance, a customer from a major beverage brand would receive the following message when using their personalisation service:

“Oops! Looks like the name you requested is not an approved one. Names may not be approved if they're potentially offensive to other people, trademarked, or celebrity names. We've worked hard to get this list right, but sometimes we mess up. If you think this is an error, please contact our Customer Care team. Otherwise, please try again, keep it fun and in the spirit of sharing!”

The list includes ‘Hitler’ and ‘Stalin,’ but not ‘Mussolini’ or ‘Bin Laden’. Many sexist and racist terms are banned in English but may not be filtered in other languages, such as their Spanish equivalents. This raises the issue of how multinational brands can handle personalisation across different cultures with varying standards for offensive content.

And what about words that don't fit neatly into defined categories but could still be harmful, particularly if they express brand criticism? A famous footwear brand learned this lesson early in the personalisation era. When the brand refused a personalisation request based on a term “sweatshop”, the refusal sparked viral backlash. The ensuing controversy surrounding the brand’s decision to deny the request, along with the customer service exchanges, ended up damaging the brand’s reputation and compromising the entire personalisation service.

This highlights the delicate balance brands must strike: while controlling associations is important, refusal to personalise based on subjective reasoning can lead to accusations of censorship, as another tech company experienced when it blocked references to political speech in its engraving service.

These situations can easily escalate, proving that what seems like a harmless term may, in fact, have far-reaching consequences.

Terms and conditions

Being upfront about the terms and conditions for personalisation may not completely eliminate misuse, but it serves the important purpose of clarifying the brand owner’s intentions, which can be crucial if issues arise. These terms should include:

• the consumer warrants that they have the necessary permissions to place the personalisation order and authorise the operator to produce the item(s) on their behalf;
• the consumer is solely responsible for the submitted content and assumes all liability for any misuse or legal consequences;
• the consumer warrants that the submitted content does not infringe third-party rights, including copyright, trade mark, right of publicity or privacy, and will not libel or defame any third party;
• the operator reserves the right to refuse any content deemed offensive, inappropriate or explicit, with full discretion, and without obligation to provide an explanation.

In-store personalisation

One idea for brand owners to consider is to shift personalisation to the in-store experience, whenever possible. This approach allows the process to be more directly managed by personnel, offering a level of control that online platforms can't always provide.

In-store personalisation helps guard against the unpredictable "imaginativeness" that often accompanies anonymous online ordering, where the lack of direct oversight can lead to unintended or inappropriate content.

By guiding the personalisation process in person, brands can ensure a more positive experience, keeping it fun and engaging, rather than risking the creation of controversial or damaging products – such as a dictator’s face on a box of morning cereals. This extra step not only protects the brand but also ensures that customers enjoy a more tailored and controlled interaction with the brand.

Marion Heathcote is a Principal with Davies Collison Cave in Syndey, Australia and Chair of the MARQUES IP Emerging Issues Team