🔴 The Day After Is Not for Tomorrow

As the political world buzzes over the appointment of the new government and, with it, a new digital minister, a paradox is becoming impossible to ignore.

The more we talk about digital technology and its problems, the less we actually do about them.

Take digital sovereignty. Everyone now agrees it is a major issue. Even the consultancy Asteres, reputed to be close to the big platforms, says so in a study commissioned by Cigref (the association of large French companies' CIOs): our dependence on Big Tech is costing us too much. Changing the model is urgent.

Since VMware, the company whose virtualization tools power the cloud, sent its license prices soaring, large French companies have been discovering, to their horror, that they do not necessarily have the budgets to keep up.

As we have often written, it's when you hit the wall that you see the wall best.

What goes for sovereignty also goes for cybersecurity. At several recent conferences, the head of ANSSI (France's national cybersecurity agency) painted an apocalyptic picture of the threats ahead of us, without the agency's usual restraint. It has to be said that over the past two years, nearly every online service has been hacked and the entire data of the French population has ended up on the dark web.

As with sovereignty, everyone knows what should be done and how to go about it. So if politicians, industry and the digital sector all share the same diagnosis, why is nothing moving?

In reality, while everyone agrees on the diagnosis, mostly because it is no longer politically possible to ignore it, nobody agrees on the solutions, nor on how fast to implement them.

Moving from "we should do" to "here is what we can do now" remains difficult. Too many people have an interest in keeping things in limbo and prolonging inaction.

For several years now, the strategy has been to sideline those who bring concrete solutions and a real will to change.

Action calls for different kinds of people, as we saw during the Covid crisis. It also challenges many entrenched powers and well-established funding structures.

And yet we will have to move from an ecosystem of problems to solve to an ecosystem of solutions to implement.

There is no shortage of examples. When it comes to digital dependence, several avenues exist:

• Replace the Windows desktops across the entire administration with Linux machines, as the gendarmerie and the air force have done.
• One could imagine switching all desktops over within a few years and putting funding strategies in place to support it.
• Even for mobile, there are two open source alternatives to Android, entirely independent of Google, that could make it possible to build new fleets of devices free of any dependence.
• One could draw up a list of replacement software and create purchasing quotas worth a few hundred million euros a year for alternative software from French SMEs.
• One could also think about what we are missing, propose the right software, and then set up a funding program for the players who want to deploy it.
• Finally, one could consider investing to shore up the open source components that are often essential to keeping the system running.

In Germany, or even in the Netherlands, foundations take care of this. In France, no such system exists, and open source players are forced to navigate the complexity of European calls for tenders to try to fund their research and development.

This transition is actually simpler than people think.

With AI, there is no longer any need to spend time recreating carbon copies of yesterday's tools: we can go straight to inventing new uses.

Even Microsoft now admits it. According to its own CEO, Excel and the classic office tools have been overtaken by the new AI services.

Why keep going into debt to fund what already belongs to the past?
One of the first measures would be to abandon proprietary formats.

With the arrival of ChatGPT, we have moved from Word files to simpler formats like Markdown, which make content easier to edit.

Why not roll these formats out across the entire administration?

Then there is the question of email. We should long ago have left this service to independent providers, of which France has many competent ones, capable of running it on their own servers.

Finally, it would have been possible to impose a directive so that the state uses, and invests its communication budgets in, only sober and reliable social networks, meaning those not known for spreading disinformation or manipulation.

It may not be very "fancy," but nobody expects public services or their staff to be cool on social media, only to be present and effective.

What is striking today is that digital sovereignty looks more and more like organic food: the demand comes from users themselves, and policymakers, overtaken by this enthusiasm, are starting to sense the risk of doing nothing.

Even if there would be a lot to say on the subject, Polytechnique's reversal of its decision to move to Microsoft, or the replacement of the director of the Health Data Hub (after taking the risk of hosting the platform at Microsoft, and several years of twists and turns), show that the lines seem to be starting to move.

I say "seem" because merely using independent tools to replicate what already exists would be a missed opportunity.

I have never believed in centralizing digital strategies. What matters today is building a shared vision, in which human beings regain their full place at the heart of the system.

For years, clever storytelling made it possible to pass off the status quo as modernization. But that cycle is coming to an end.

AI is such a major revolution, and its impacts are so unpredictable, that we have to prepare.

In today's world, any French citizen can download government documents, the budget of their region or their city, have them analyzed by AI and then ask extremely precise questions.

The same tools let citizens find out how to fill in everyday forms and reach, for the length of a session, the skill level of an experienced civil servant. Obviously, AI does not always have the perfect answer, but for many French people, the services they use every day are going to start showing their limits.

The era of the augmented citizen requires high-level public services, but also augmented civil servants, given the latitude they need to operate at the new speed now expected of them.

Yet what we see is an administration still stuck in the 2010s, the years of "nice" Big Tech and of Web 2.0, when open data strategies and digital benevolence could still be sold to citizens.

That was before AI indiscriminately swallowed our digital government to build data models that hope to replace it.

While every politician has their eyes fixed on 2027, few have understood that citizens' skill level is about to explode: boosted by AI, they will learn at an accelerated pace and with a level of expectation never demanded before.

An entire generation of public officials has not understood what is at stake with AI, namely that a reversal of power in favor of citizens, who are far more agile, better informed and more demanding thanks to AI, has already begun.

And sadly, many remain convinced that it is enough to talk about a sovereign future that everyone wants but that they do not know how to deliver, which also masks their inability to imagine what comes next.

If only they knew.

For paying subscribers, I share my daily monitoring on a private channel reserved for you; here is the link to access it.