FIMI? Maybe it’s time to revise the concept

or simply drop it alltogether

                                                    (thanks Eva Bauchau for proofreading and additions)

The concept of Foreign Interference and Manipulation of Information, or just FIMI, gained traction in recent years thanks to the activism of European External Action Service (EEAS) who works tirelessly to make this notion accepted and utilized by organisations researching in the field of disinformation.

However, since the introduction of the Digital Services Act (DSA), the lessons from the Covid Crisis start to get written. Now, we are two years in since Putin’s war of aggression against Ukraine and the European information sphere. Now, should be the time to revise this notion.

The first question that springs to mind is: what is FIMI as a concept? Is there a proper definition out there that would make the concept useful, and what does this definition bring to the fight against disinformation.

Foreign Interference and Manipulation of Information, what is is about?

Here is how the EEAS defines it:

“A mostly non-illegal pattern of behaviour that threatens or has the potential to negatively impact values, procedures and political processes. Such activity is manipulative in character, conducted in an intentional and coordinated manner. Actors of such activity can be state or non-state actors, including their proxies inside and outside of their own territory.”

This definition already poses some challenges, especially when one tries to put it into practice. The definition claims to focus on “patterns of behaviour”, but the assessment of “negatively impacting values” is impossible to determine by behavioural criteria. “Values” and anything related to it, has nothing to do with AI created profile pictures, fake accounts and inauthentic number of retweets.

For instance, reading Margot Fulde Hardy and Camille François’s attempt to establish a workable framework for studying and analysing FIMI through behavioural criteria, one can easily realise that no TTP or any other form of criteria can be used to assess “values.” This becomes even more problematic when trying to assess what negatively impacts them.

And this is not just about a lack of criteria. The core problem is who is supposed to define the ‘values’ ? We are talking about a notion, FIMI, that is framing policies that are on the fringe of free speech regulation. And FIMI is targeting “mostly non illegal patterns of behaviour”. So, if the criteria is not a legal one, and if non illegal behaviour can become the target of anti-FIMI policies, what are the values that constitute an actionable justification for EU policies against “FIMI”?

Real cases of Disinformation shows that the attackers (let’s call them “The Russians”) aims to play on the divisive subjects of our democratic societies. “The Russians” will precisely target the “values” that do not reach consensus. Support for Israel military intervention in Gaza for instance, or Europe policy against migrants . these are the likely targets of disinformation attacks, are we sure “FIMI” has the mandate to establish what are our values and what are the non-illegal behaviour that negatively impacts them on these topics?

Let’s examine the very same issue on a more tangible example. “A mostly non-illegal pattern of behaviour that threatens or has the potential to negatively impact the political processes”. It is indeed possible to think of behavioural criteria that could be used to detect an attempt to negatively impact a political process. We can think of a Facebook pages changing names just before an election, or inauthentic accounts and websites impersonating actual political parties.

Yet, the line between interfering in an election process and preventing interference in an election process can be blurry. Accusing a domestic party to be playing in the hands of a foreign power is an all times classic, especially as fascist and populist ideologies are on the rise in Europe. In such cases, the FIMI notion, mixing in its definition “state or non-state, including proxies inside and outside their own territory” can easily be used to accuse a domestic political opponent of being the puppet of a foreign power.

Foreign for a reign

To really prevent the FIMI notion from being used in national populist propaganda we should rethink it, so as to address this core problem: what is “foreign” and what is “domestic”? Is it really necessary to establish a distinction between the two and if so, where to draw the line? We know that disinformation aims at playing on internal divisions and blurring the borders between what is domestic and what is “foreign”, why choose a definition that plays into their hands?

Let’s look at the issue under a different angle. A national formerly recognised newspaper is bought by a disinformation actor, and subsequently publishes conspiracy theories and state-sponsored propaganda, some directly copy pasted from Russian false narratives. It employs journalists with a press card, pays taxes as if it was a real newspaper and behaves exactly and legally as if it were the former publication, well-known and respected. Is it “Foreign”? (This example is an actual newspaper called France-Soir)

What do we do with RT France, RT Spain, RT Germany which are legally declared companies in France, Spain or Germany, employ local journalists and issue content in local languages? What criteria do we use to call them “foreign”, and what do we do so these criteria will not be turned against CNN or the BBC? At the expense of stating the obvious, the BBC and Russia Today are not in the same business, so we need a definition of “foreign state media” that can target the one and spare the other. If each time we have to redefine, reinvent, distort and twist the FIMI notion so that it fits the reality of the disinformation field, maybe it is high time we dropped this notion altogether.

Definition(s)

For High Representative / Vice President Josep Borrell in its foreword introducing “2^nd EEAS report on FIMI threats”, FIMI is:

Foreign actors, who engage in intentional, strategic and coordinated attempts to manipulate facts, to confuse, sow division, fear and hatred. FIMI is closely connected to both hybrid threats and cyber threats and has become a crucial component of modern-day warfare.

The notion, who was already about values and political process, is now coupled with cyber threats and modern-day warfare. The focus switched from “pattern of behaviour” to the intention to “confuse, sow division, fear and hatred”, intentions that can’t be measured or assessed by behavioural criteria.

Borrell’s foreword directly contradicts the “LGBTQI+ FIMI” definition that was set in the FIMI report on LGBTQI+ communities. In this report it was essential to separate FIMI from other illegal and harmful behaviour such as hate speech.

“This first subsection provides clarification on some of the key terms used in this field of study. The aim is to establish a clear distinction between FIMI and other harmful behaviours such as hate speech, harassment, hate crimes, disinformation, and gendered disinformation in order to provide a clear concept of LGBTIQ+ foreign information manipulation and interference (FIMI).”

Although the 2^nd FIMI report claims to have collected and investigated 750 “FIMI Incidents”, criteria for selecting or defining what is a “FIMI incident” are nowhere to be found. In the report one can even find changing definition of FIMI. Sometime FIMI seems to be an all mighty threats on anything and everything

“Foreign Information Manipulation and Interference (FIMI) is a global challenge – one that knows no geographical or thematic borders. Anyone can be the target of information manipulation.”

Elsewhere in the same report one can find that

“Compared to cyber, FIMI has a stronger socio-cognitive component that also plays a role in the analysis and the design of responses beyond the technical dimension”

… which seem to invoke notions of Social Cognitive Theory?

How can we aim toward a whole-of-society approach on FIMI if even the EEAS who designed the notion changes the definition from one report to another, and even in the same report?

Framing FIMI in International Law

Seemingly aware of the problems, the EEAS looked for outside Europe solutions, and ordered a “study for international norms on FIMI” to Oxford university in an attempt to elaborate a working definition for the FIMI concept. The Oxford team then tried to answer two main questions: where does the FIMI concept stands alongside the other concepts (i.e. influence operations, disinformation etc); and can the FIMI concept be framed by international law.

The first question lets appear that the FIMI concept is just one more concept, adding up to a pile of preexisting definitions. The downside of adding a new concept, together with a new definition, is that it goes against fighting disinformation all together. We have to make things simpler not more complicated. Disinformation strategists are already trying to burry us into a pile of disinformation items to fact check¹ . Adding new layers of concepts and definitions and trying to impose new top-down standards disconnected from the field plays into that strategy.

How do we fight conspiracy theories if even ourselves in the disinfo industry, have to be “connecting the dots”, “reading between the lines” and “doing our own research” between dozens of different concepts? How do we speak about transparency if notions and criteria are buried under three, four, five, six different overlapping definitions and competing bodies?

Every layer makes the subject more obscure, more complicated, more difficult to apprehend to regular EU citizens, the very people we want to protect from disinformation. We can’t promote transparency and values of truth by one hand, and by the other hand overcomplicate the matter. How can we drive media literacy to help people navigate the hurdles of mis- and disinfo, and complexify the notion so much that even we, as experts, are unsure of what it actually means?

The second framework that Oxford tries to apply to the FIMI definition in its study, is international law. One may think that the FIMI legal framing should include the most advanced and recent European regulation on the subject: the Digital Services Act. However, that is not the case. The Oxford report on the application of FIMI within the framework of international law does not even mention the DSA. Hence, the FIMI notion seems impractical to anyone in need to connect with the work of the DSA teams.

The FIMI notion was conceived at a time the DSA was not yet existing and it was not revised since. FIMI is thus out of the methodology and out of the tools of the world’s most advanced regulation on disinformation, which can become problematic for people working in the field, and who want to connect with the DSA teams, and need to align their works on DSA standards and templates.

Behaviour versus Narratives

Of course, the Oxford team examining the FIMI notion had to address one of the main operational problems that we are all facing: the frontier between behaviour-based methodology and narrative-based methodology.

As stated before, behaviour-based methodology looks at the accounts’ behaviours: the number of likes, followers, time of posting, inauthentic profile pictures etc. Narrative-based methodology, however, looks at the illegal content, hate speech, conspiracy theories and other type of fringe or illegal activity.

Both approaches have their pros and cons. Looking at behaviours spares one from addressing the content and avoids problematics of free speech and censorships. But assessing what is authentic behaviour and what is inauthentic is also a complicated business. The FIMI notion initially claimed to be behaviour-centered, but, as the Oxford teams puts it:

“This in turn means that an analysis of the content in question is inescapable. To know if States may prohibit, take down or otherwise limit a certain information operation, including strictly lawful speech acts falling within the concept of FIMI, one needs to assess if the content in question amounts to war propaganda, incitement, defamation or affects other public policy interests, such as public health, security or order.”

Then again, one of the core ideas of the behaviour-centric notion of FIMI was precisely to avoid involving the States in assessing what content is allowed and what is not. The FIMI notion was elaborated to avoid “policing” content shared in online spaces, and yet it is faced with inextricable need to assess content to legitimize that the content falls under the FIMI umbrella.

In a new European political landscape with raising populist and nationalist far right forces, how much do we trust regulations of war propaganda or anti-public policies on security and order, sanctioned by the EEAS’ FIMI notion?

Signals and Shadows

1. See for Instance Operation Overload https://checkfirst.network/wp-content/uploads/2024/06/Operation_Overload_WEB.pdf