I need your clothes, your boots and your motorcycle
Until very recently, the concept that machines could identify and respond to images was as far-fetched as a cyborg being sent back in time to kill a future leader. And the first time you hear of the capabilities of some of the newest “wearable” consumer products they seem almost like science-fiction.
One example is Google Glass, a pair of internet-connected spectacles, which combine visual recognition software with voice and gesture control. The glasses use cameras to identify what the wearer can see and then to display related electronic images in the field of vision. For example, if the wearer looks at the Eiffel Tower then the Wikipedia page for the landmark might pop up.
One of the most promising applications for the technology is almost instantaneous language translation. As well as being able to convert audio Babel fish-style, so the wearer can converse in multiple foreign tongues, software such as Word Lens can read and translate text and signs before superimposing the translation in the place of the original words. Perhaps most important for any cutting-edge technology, Google Glass also allows you to take selfies and watch cat videos.
Whilst the technology may be ready, it is clear that the aesthetics have a way to go. Wearing a clunky box on the side of your glasses is currently about as cool as a Bluetooth earpiece. Fortunately, Google seems to have something else in the pipeline. A number of pending patent applications and the recent announcement of its “smart contact lens” project indicate an option more fashion conscious technophiles may soon have an alternative option. Their less intrusive nature may at least make the technology more appealing to mainstream consumers.
This augmented reality, or “Terminator-vision”, also opens up a new world of potential advertising possibilities, some of which we are likely to have already been unwittingly exposed. The ability of consumers to easily skip advertisements, means that product placement is increasingly important. Using digital replacement technology, brand names and even entire objects can be placed or removed from films and recorded tv to suit specific markets.
Going one step further, digital replacement technology is now adopted at live sports events. The content of billboards around the playing area is captured in the usual way by the on-site cameras but before the footage reaches the armchair fans the signal is altered so that the billboard content is replaced with market-specific advertisements. This means the content received overseas is sensitive to local custom, for example those which prohibit gambling. It also means the advertising space can be carved up into many more segments and replaced with content relevant to the audience.
Combining all of this with “deep-learning” technology which can recognise concepts, such as whether a YouTube video is of a cat, brandowners might seek to gain an advantage by paying search engines to display an advertisement whenever the “Go-Go-Gadget spectacles” see an object, whether that is a bunch of flowers, a watch or a handbag. One satirical interpretation of this is ADmented Reality which claims to be a “slightly more realistic version of Google’s augmented reality glasses – now featuring contextual Google Ads!“.
A savvy search engine might use the technology to identify logos in the wearer’s field of vision and sell advertising space adjacent to or even in place of real-world logos. Is that Rolex watch really a Rolex when what the wearer sees is an augmented “Fauxlex” in the place of the official brand name? Is there any guidance as to how the courts would treat such activity?
Standing on the shoulders of giants
Surprisingly, no single part of the Google Glass product is revolutionary. It is really “just” a digital chimera composed of several ubiquitous functions of everyday smartphones: hands free; wi-fi; image search and so on. To that extent then, it seems natural to consider whether the existing case law on keyword advertising brings any clarity to this vision of the future. The test for trade mark infringement remains the same but the update with this iteration is that instead of words being used as triggers for advertisements it is images.
In the European Union, the test for trade mark infringement under Article 5(1)(a) of the Harmonisation Directive and the interpreting case law requires:
- use in the course of trade;
- of an identical sign;
- in the relevant territory;
- in relation to identical goods or services for which the mark is registered;
- without the consent of the proprietor; and
- the use must be liable to have an effect on one of the functions of the trade mark; and
- the origin function of a trade mark is adversely affected if the advertisement does not enable the average internet user, or enables them only with difficulty, to ascertain whether the advertised goods or services originate from the trade mark proprietor or from a third party.
Article 5(2) of the Directive gives enhanced protection to marks with a reputation. Infringement occurs where consumers make a link between the sign used and the registered mark and the advertiser either takes unfair advantage of, or causes detriment to, the distinctive character or reputation of the mark. After considering this provision in L’Oréal and eBay, the Court of Justice stated that a claim for detriment to the distinctive character of a mark, or “dilution”, will succeed where reasonably observant internet users cannot distinguish whether the advertisement originates from the brand owner or a third party. In doing so it neatly aligned this part of the test under Article 5(2) with that in Article 5(1)(a) and Article 5(1)(b) such that a likelihood of confusion as to origin will always be required.
In addition, Article 6 of the E-Commerce Directive requires that electronic advertisements clearly identify i) that they are in fact commercial communications and ii) the person or business on whose behalf they are made.
Firstly, even if the search engine was selling “key images”, the display of which will trigger an advertisement, it is unlikely that this activity will constitute trade mark infringement for the operator. In Google France the Court of Justice said that even though Google sold advertising space through its AdWords service on the basis of the trade marks, it did not use the trade marks in its own commercial communications. Instead, Google allows its clients to use the trade marks without itself using those signs. The fact of creating the technical conditions necessary for the use of a sign and being paid for that service does not mean that the party offering the service itself uses the sign “in the course of trade”.
Unlike the advertisers, it therefore appears that the search engines have a good argument that their “internal” use of keyimages will not infringe trade marks. Nevertheless, they retain obligations under the E-Commerce Directive to ensure that consumers recognise which content is advertising and who is responsible for it. The main question for search engines and advertisers is therefore whether consumers are likely to recognise that the advertisement is indeed an advertisement or some other “generic” content displayed by one of the services “embedded” in the product.
On a desktop, the results results page usually indicates an advertisement by the use of the word “Ad” and the use of a different font or background. In the world of augmented reality, if the image is replaced seamlessly, such as with billboards in tv broadcasts, then perhaps the context itself will be sufficient to make it clear that it is an advertisement. The augmented image is in the identical place of a real-world advertisement after all.
As an alternative, when the keyimage is not digitally replaced, an advertisement might pop up as a balloon on the screen adjacent to the trigger. To distinguish that content from more useful information, such as navigation directions, it could state simply “Ad”.
Overall this may be too intrusive for the technology to be attractive to mainstream consumers but there is clearly a balance to be struck between capturing the wearer’s attention and making the advertisement non-intrusive. It will be interesting to see what methods the search engines adopt and which will prevail. Like with smartphone apps, perhaps we will see ad-free versions of the technology available for a premium.
Brands which choose to use keyimage triggers must ensure that the resulting content clearly indicates that it is an advertisement. If the keyimage trigger is a logo of another brand rather than a generic image, the advertisement must also ensure that consumers recognise that the content is unrelated to the brandowner.
In Interflora, the courts emphasised that whether there is confusion will depend on the identity of the parties, the perception of the targeted public as well as the nature of the advertisement and context in which it is displayed. Just like with desktop search, as the demographic evolves from “noobs” to a tech-savvy demographic, the threshold for a likelihood of confusion will change.
In the UK at least, possibly the best weapon in a brandowner’s arsenal may be the copyright and design right in its logo. With copyright, as there is no “fair-use” doctrine, any reproduction of a substantial part of the logo will infringe unless it falls into one of the narrow, and generally non-commercial, exceptions. For design right, anything which creates the same overall impression will infringe provided the design is new and has individual character. That means that the use of an identical or similar image by a search engine or by an advertiser as a trigger to further their own business purposes is likely to infringe. Keywords and generic images which are not protected by copyright or design right will not be affected.
There’s a storm coming
Whilst businesses can in principle use a competitor’s word trade mark to improve search engine results rankings, the battlefield over image trade marks seems to favour those brands with unique logos. Who succeeds, only time will tell, but the machines seem certain to play a part for some time.