Visual Search Engine
Arriba Soft (now Ditto.com) operates a visual search engine using pictures instead
of standard text result set. Arriba's bot crawls the web and stores in its database
reduced size thumbnails images of the pictures found. Consequently, when a user
performs a search for photographer, such as Leslie Kelly, the search engine will
return a small thumbnail image of a western barn taken by the photographer Leslie
Kelly. Users are then linked to the originating web site on which the picture is
located.
In January 1999, Arriba's bot crawled the website belonging to the aforementioned
photographer Leslie Kelly and processed 35 photographs into its database. When Kelly
discovered his images on the Arriba search page, he complained to Arriba, who promptly
removed the offending thumbnail images and put his website on its "do not crawl"
list. Subsequently, Kelly found additional photographs of his on the Arriba search
page that the Arriba bot had obtained from other websites. Kelly complained, and
Arriba again removed the offending thumbnail images and put the sources websites
on the bot's "do not crawl" list.
This was not sufficient for Kelly, and he subsequently sued Arriba for copyright
infringement (for reproduction and display). Arriba claimed fair use.
The Fair Use Test
In order to determine whether Arriba Soft has a valid defense, the court needs to perform the four-part Fair Use Test.
While each court appears to apply the test in its own way, a particularly lucid application was described in the
case of Field v. Google. In this case we will use the court's guidance to map to
an actual Fair Use Score using our Fair Use Visualizer.
1 - Purpose of Use
The first element actually has three sub-elements:
Use. This sub-element looks at how the work is used on the spectrum of commercial use to non-commercial use. The
less commercial a use, the more a fair use finding is favored.
Arriba’s use of Kelly’s images was commercial, although it was more incidental and less exploitative
in nature than more traditional types of commercial use. Arriba was not directly
profiting from Kelly’s images, as they were among thousands of images in the search
engine database. Because the use was not highly exploitative, the court found that the commercial nature
weighs only slightly against a finding of fair use. Since its slightly against, we'll give this sub-element
a Visualizer score of 45 (slightly less than the inflection point of 50).
Purpose. This sub-element looks at the statutory ennumerated favored purposes for using the work, such
as news reporting, commentary, criticism, paraody, etc. There is no indication that any of these purposes
where at issue here, so we'll give this sub-element a Visualizer score of 0.
Transformation. This sub-element looks at whether you are just using the work because you are too lazy to
do it yourself (not favored), or whether you are bringing something new to the table (favored). Although
Arriba made exact replicaions of Kelly’s images (not favored), the thumbnails were much smaller,
lower-resolution images that served an entirely different function than Kelly’s original images (favored).
The court focused on the fact that the use of the thumbnails involved more than
a retransmission of Kelly’s images in a different medium. Kelly’s use is aesthetic,
whereas Arriba’s use is utilitarian. It is unlikely people would use the low-resolution images for aesthetic purposes.
In un Nezez v. Caribbean International News Corp., the copying of a photgraph that
was intended to bu used in a modeling portfolio and using it instead in a news article
was held to be a transformative use. By putting a copy of the photograph in the newspaper,
the work was transformed into news, creating a new meaning and purpose for the work.
The court felt that the use of Kelly’s images in Arriba’s search engine is was analogous to the situation
in un Nezez because Arriba has created a new purpose for the images.
The thumbnails do not stifle artistic creativity because they are not used for illustrative
or artistic purposes and therefore do not supplant the need for the originals. In
addition, they benefit the public by enhancing information-gathering techniques
on the internet. This sub-element gets a Visualizer score of 90.
2 - Nature of Work
Creative works are more protectable than fact-based works. However, published works
favor fair use since the "first appearance of the artist’s expression has already
occurred." I don’t really know what that means, but in this case the court said
that since Kelly’s images appeared on the internet before Arriba used them in its
search engine, that Kelly would only get a slight advantage for this element. Since its
slightly against, we'll give this element a Visualizer score of 45 (slightly less than the
inflection point of 50).
3 - Relative Amount
Like the first element, this element also consists of three sub-elements:
- Amount Taken
- Quality
- Ratio of Amount Taken to New Work
The trade-off here is: 1) using all of the work weighs against you; and 2) using
only as much as you need, even if it is all of the work, weighs for you. Arriba
used all of the work (the whole images), but that is what was needed in order to
accomplish its purpose. The court did not go into the individual sub-elements, but found that
this element was neutral to both parties. This will map to a Visualizer score for the sub-elements as
follows: taken = 0, quality = 100, ratio = 50.
4 - Market Effect
This element consists of two sub-elements:
- Market Effect for the Original
- Market Effect for Derivative Works
In this case, we need to look two things: 1) whether Arriba’s use would
result in a substantially adverse impact on the potential market for the original
works of Kelly, and 2) whether there would likewise be an adverse impact on derivative
works.
Arriba’s use of Kelly’s images in its thumbnails does not harm the market for Kelly’s
images or the value of the images. By showing the thumbnails, Arriba guides users
to Kelly’s website, not away from it. The thumbnails are not a substitute for the
full-sized images because they lose their clarity when enlarged. If the user wants
a quality image, they would have to visit Kelly’s website.
Arriba’s use would also have little effect on Kelly’s ability to sell or license
his images, as Arriba does not sell or license its thumbnails to third parties.
This court found that this factor weighs in favor of Arriba. This will map to a Visualizer
score for the sub-elements as follows: original = 75, derivative = 75.
Fair Use for Thumbnails
In the end, the court noted that of the four elements of the Fair Use Test, two
weighed in favor of Arriba, one weighed slightly in favor of Kelly, and one was
neutral, and therefore Arriba wins.
While this reconciliation of the elements is somewhat linear and rudimentary, the
advanced AI of our Fair Use Visualizer
indicates a Fair Use Score of 58, confirming that the court did in fact reach the correct conclusion.