Under the striking glass vaults of the Grand Palais in Paris, world leaders and tech titans gathered for a two-day summit dedicated to artificial intelligence. The heart of the event space, tucked in between the Seine and the iconic Champs-Élysées, was organised like an exhibition hall packed with dozens of stands showcasing the various ways AI can be used.
Climbing up the stairs in the building’s majestic nave, a display of flags decorated the railing, boasting the different nationalities taking part in the summit. At the top, a handful of key players gathered on Monday afternoon in the VIP lounge to discuss the impact AI could have on democracies.
But a few VIPs were missing. Representatives from some of the most influential companies in the sector including Open AI, Google and Microsoft played hooky.
Their absences were criticised by Meredith Whittaker, president of the end-to-end encrypted messaging app Signal who worked at Google for over a decade and helped organise mass walkouts, partly fuelled by the company’s handling of AI ethics, before resigning in July 2019.
“The large-scale approach to AI is damaging [societies],” she said.
Security threats
“AI has profound consequences on our private lives,” Whittaker continued. To illustrate her point, Whittaker referred to the handful of telecoms companies in the US including Verizon and AT&T who fell victim to a sweeping Chinese-linked espionage operation known as Salt Typhoon in December last year. A vast amount of US users had their metadata taken as a result, and officials associated with both Kamala Harris and Donald Trump’s campaign ahead of the November presidential elections were targeted.
“From a security standpoint, what happened was a disaster. And it was made possible by backdoors [programmes that allow hackers to access a computer system or encrypted data remotely] being installed by authorities who were meant to be the only ones to access them,” the Signal president explained.
For Whittaker, AI was “born from the business model of surveillance”. And when it comes to hacking, “data can be used as a weapon” against US democracy.
The threat is undeniable for Latvian President Edgars Rinkevics, who shares Whittaker’s views. Rinkevics explained how, “due to the geopolitical and geographical situation” of his country, “Latvia has been exposed to cyber-attacks and disinformation campaigns” in recent years. With Russia as its neighbour, Latvia and the other Baltic countries Estonia and Lithuania bore the brunt of cyber-attacks in recent years. EU member states have repeatedly been the target of Russian cyber-attacks in 2024.
Before he went into detail about the specific threats Latvia faces, Rinkevics stressed that “the priority is to protect critical infrastructure and monitor the situation in the Baltic Sea”, where several undersea telecoms, power and data cables were sabotaged in recent months.
“With regards to democracy, AI has been used to meddle in elections, especially in Romania,” Rinkevics added. “In Latvia, we passed a law that requires [all content created by] AI to be labelled” so that people can identify when it has been used. Rather than being wary of the technology, Rinkevics believes “we need more expertise to understand exactly where cyber-attacks come from and how to better protect our democracies”.
Read more‘Aggressive’ Russian cyber attacks boosted Romania’s pro-Moscow presidential candidate
In a leap of faith, Latvia signed an agreement with Microsoft in December last year to develop a National Center for Artificial Intelligence. Its aims include promoting AI and other digital solutions to modernise the country’s administrative processes.
Need for global AI safeguards
From eerily accurate deep-fakes to social media accounts usurped by scammers for money, artificial intelligence has many faces. And that is partly why it is such a threat to democracies. “AI can make it easier to carry out cyber-attacks, which have become more sophisticated now that tools generating complex code can be used to this end,” warned Marie-Laure Denis, President of the French Data Protection Authority (CNIL).
Denis believes that solutions to protect democracies should stem from a diverse range of safeguards, starting with the general data protection regulation (GDPR), implemented in the EU since 2018 that sets out guidelines on how to collect and process personal information from people both within and outside of the continent. “We should develop a trusted AI tool to strengthen the protection of our rights,” she added. “Without guarantees, we can’t have trust. And without trust, we can’t develop AI in the long-run.”
But when it comes to AI and democracy, not all countries feel equally included in the conversation on regulations. Pakistani lawyer and online rights activist Nighat Dad intervened to point out global disparities. “Are our exchanges at this roundtable democratic? Are we talking about the whole world or just the democracies of the Global North?” Dad asked. She founded the Digital Rights Foundation in 2012, an NGO focused on protecting human rights defenders in digital spaces. “Moving forward without diverse voices is unimaginable. All democracies must be able to express themselves.”
Mathias Cormann, secretary-general of the OECD, echoed Dad’s remarks when he spoke of the need for “more effective international cooperation on AI”.
“We need a generalised governance framework to help us develop safe AI,” he said. “But right now, we are a long way from that.”
“At this stage, we don’t know exactly where we’re headed when it comes to AI,” Latvian President Rinkevics concluded. “The priority for companies is to reap the benefits [of AI]. And the priority for countries is likely a little different. But in any case, we are a long way from having a global AI agreement.”
This article was translated from the original in French by Lara Bullens.
Leave a Comment