Add initial motivations section

This isn't really saying anything new that the explainer isn't saying
already. Adding this so that people can get an idea of the goals of this
spec without having to read other sources.

docs/spec.bs

@@ -25,7 +25,20 @@ Web Environment Integrity is best suited for detecting deceptive [=web environme
 
 ## Motivations ## {#motivations}
 
-<i>Todo</i>
+<i>This section is non-normative.</i>
+
+Users often depend on websites being able to establish the trustworthiness of the
+[=web environment=]. When users are playing online games for instance, they are trusting that other
+users are not cheating. Or when they are browsing social media websites, they are trusting that
+other users are not faking engagement to make posts popular.
+
+Websites currently need to establish this trust relationship without any support from
+[=user agents=]. This can result in websites collecting signals that can be used to fingerprint
+users.
+
+<dfn for="Web Environment Integrity">Web Environment Integrity</dfn> is being proposed as a privacy
+preserving API to provide low [=entropy=] trust signals for the [=web environment=]. This API relies
+on [=user agents=] having access to an [=attester=].
 
 ## Examples ## {#examples}
 
@@ -60,6 +73,11 @@ The <dfn for="web environment">web environment</dfn> is defined as <i>TODO</i>
 
 The <dfn for="content binding">content binding</dfn> is defined as <i>TODO</i>
 
+<dfn for="entropy">Entropy</dfn> in this specification refers to Shannon entropy. This is the amount
+information conveyed.
+
+# Attesters # {#attester}
+
 The <dfn for="relying party">relying party</dfn> is defined as any party that relies on the Web Environment Integrity API
 outside of the [=user agent=].