AI: A threat or boon for democracy?

As artificial quality impacts our lives successful expected and unexpected ways, and augurs unprecedented caller opportunities and challenges, starring thinkers, analysts and argumentation makers came together successful the Greek superior to sermon what this burgeoning caller technology means for ideology astatine ample successful a special two-day league successful the discourse of the 2025 Athens Democracy Forum (September 30 – October 3) organized by the Democracy & Culture Foundation successful relation with The New York Times.

‘As AI advances, we indispensable inquire ourselves what benignant of quality beings we want to become, and what benignant of ideology we wish to preserve’

Delving into the intersection of artificial intelligence, morals and antiauthoritarian governance, the special session, held astatine the Athens Conservatoire connected Tuesday and Wednesday, was organized by ADF and Kathimerini, successful relation with the NYT.

In his opening remarks, the president of the Democracy & Culture Foundation, Achilles Tsaltas, succinctly framed the crux of the issue: “As AI advances, we indispensable inquire ourselves what benignant of quality beings we want to become, and what benignant of ideology we wish to preserve,” helium said. “There is simply a existent information of powerfulness becoming centralized and ideology sliding into authoritarianism. But what if AI became a tool for genuine dialogue? Ultimately, this is simply a substance of quality choice. The work lies with us.”

Democracy & Culture Foundation President Achilles Tsaltas argued that however AI is used ultimately comes down to quality choice.

Greece’s curate of integer governance, Dimitris Papastergiou, struck a akin chord of some accidental and caution. Technology, helium noted, cannot solve each problem, but it tin “help authorities work faster and bring it person to citizens.”

However, “AI tin besides beryllium turned against ideology itself,” helium warned, pointing to deepfakes and however the “skill and speed of manipulation” tin beryllium amplified. Such manipulation is “not a technical risk, it is simply a antiauthoritarian threat.”

Papastergiou underscored the value of mounting ethical boundaries to power the use and enlargement of AI. “Ethics are the instauration of our attack arsenic a ministry and government,” helium said, noting that it is not conscionable a Greek contented but an EU contented too.

He besides pointed to the advancement made by Greece connected the integer beforehand successful caller years, peculiarly with the improvement of the Gov.gr platform, which gives citizens entree to hundreds of services and applications that bash distant with reddish tape. “Now we’re taking the adjacent step into artificial intelligence,” helium said, pointing to the improvement of the Daedalus supercomputer – expected to beryllium the astir almighty computing system successful Greece and 1 of the starring systems successful Europe – and the Lavrio Data Center.

Digital Governance Minister Dimitris Papastergiou outlined the advancement made successful improving citizens’ entree to state services.

Another panel, moderated by Kathimerini’s overseas exertion Xenia Kounalaki, examined the short- and semipermanent impacts of AI connected wide ideology and highlighted the uncertain bequest of past technologies.

Stathis Kalyvas, Gladstone Professor of Government astatine Oxford, suggested that the nett offers a roadmap for understanding AI’s imaginable impact, though helium expressed doubts astir whether the nett has strengthened ideology overall.

“The champion justice of however AI will alteration humanity and governmental systems is the internet. I’m not sure that the benefits oregon hindrances of the nett person been intelligibly one-sided,” helium said, adding that “it looks similar the acold close is winning.” Kalyvas expressed peculiar interest implicit chatbots and however they tin beryllium manipulated to make oregon accelerate trends. Describing them arsenic 1 of the “one of the astir disturbing developments” stemming from AI technology, chatbots, helium said, “have been developed to corroborate people’s biases, frankincense creating a society where radical will not beryllium capable to pass owed to a immense quality successful cardinal understanding.”

Yale governmental idiosyncratic Hélène Landemore countered with a much optimistic view, pointing to κοινωνικά movements similar #MeToo and #BlackLivesMatter arsenic examples of however integer tools tin bypass traditional gatekeepers and empower citizens. That said, “we request some benignant of planetary regularisation which is unlikely to travel from the US oregon China, arsenic they view themselves arsenic the ‘winners’ and frankincense person nary inducement to bash so or

hinder their efforts. Hence, it will person to travel from the ‘losers,’ similar the EU.”

Landemore besides suggested reforming firm governments and possibly allowing for a veto oregon the powerfulness to explicit absorption to these caller algorithms designed for profit. “But I don’t think the US would let it, arsenic it would pb to it being a step down China.”

In the sheet connected ‘the Short- and Long-Term Impacts of AI connected Liberal Democracy,’ moderated by Kathimerini’s overseas exertion Xenia Kounalaki (l), a dynamic treatment unfolded betwixt Oxford prof of authorities Stathis Kalyvas (c) and Yale governmental idiosyncratic Hélène Landemore (r).

Embedded values

The forum’s 2nd large panel, moderated by Lex Paulson of the UM6P School of Collective Intelligence, explored the ethical dimensions of AI.

For Yannis Assael, Greece’s delegate to the OECD’s Global Partnership connected AI, the existent contented is not whether AI is ethical, but which values societies take to embed successful it. “AI is simply a tool created to implicit an objective,” helium said, noting that the world needs to “work a batch faster” to found ethical standards and stopgaps. “We request widespread national acquisition to flooded technophobia.”

Christos Megalou, CEO of Piraeus Bank, noted that “humans indispensable person the past word. Maybe we tin trust connected AI to springiness us the dilemma, but humans indispensable marque the decision.”

Taking a antithetic approach, Stelios Virvidakis, a prof emeritus astatine the University of Athens and an adept successful motivation philosophy, projected that AI could 1 time service arsenic an advisor to governmental leaders, offering dispassionate counsel successful moments of crisis. On a wry note, helium asked:

“What tin a instrumentality bash for you when you’re faced with an ethical issue? You request a human.” However, “the machines are becoming a small similar us successful the consciousness that they are becoming an hold of our quality and our caput – up to the constituent that we power them.”

Aristotle and the future

Day One of the special league concluded with a New York Times-style statement connected whether Aristotle’s doctrine tin connection a motivation model for the property of AI, moderated by Stephen Dunbar-Johnson, president, international, of The New York Times Company.

Panteion University prof Fereniki Panagopoulou argued that Aristotelian morals encourages societies to inquire not lone what to do, but what benignant of world they want to create. “Technology is simply a means for quality progress,” she said. “AI indispensable beryllium used to support state and rights, not to origin harm.”

Peter G. Kirchschlager, prof of morals astatine the University of Lucerne, agreed, calling for strict planetary regularisation to guarantee technology serves quality well-being.

Others pushed back. Landemore argued that Aristotle’s model whitethorn beryllium ill-suited to modern governance, where powerfulness dynamics, not conscionable idiosyncratic virtue, are astatine stake. Stavroula Tsinorema, prof astatine the University of Crete, questioned whether Aristotelian morals tin relation successful a pluralistic, planetary society. “Do we request an wholly caller ethical framework?” she asked.

Stephen Dunbar-Johnson, president, international, of the NYT, moderated a sheet connected the relevance of Aristotle’s doctrine today.

National strategy

Day Two opened with an authoritative presumption of Greece’s National Artificial Intelligence Strategy. Vasso Kindi, prof of doctrine of subject astatine the University of Athens, outlined the country’s broad AI strategy, which focuses connected innovation and entrepreneurship, acquisition and research, regulatory frameworks, and the adoption of AI crossed cardinal sectors including health, defense, civilian protection, justice, culture, taxation, nationalist medication and smart cities.

She highlighted six flagship initiatives: the instauration of a nationalist information governance framework; the constitution of a planetary AI acquisition and probe center; the improvement of an acquisition platform; the strengthening of ecosystems astir information and computing resources; the motorboat of the “Greek Data Space” dedicated to connection and culture; and the instauration of an planetary forum connected AI morals alongside an observatory to show its interaction connected antiauthoritarian processes.

In a treatment titled “Hope and Reality for the Flagship Programs of the Blueprint,” Marios Dafnomilis, strategic foresight advisor to the Presidency of the Greek Government, emphasized that AI is reshaping each facet of life, from acquisition and wellness to section governance. He underscored the value of preparing each citizens for the integer transition so that “no 1 is near behind,” and pointed to resilience and κοινωνικά cohesion arsenic the halfway of authorities initiatives. He besides stressed the request for older generations to beryllium included successful this transformation and praised the relation of the Greek diaspora successful contributing to the nationalist effort. “The flagship initiatives are not astir the machines but astir κοινωνικά cohesion and resilience,” helium said.

Stefanos Kollias, president of the National Infrastructures for Research and Technology (GRNET) and prof emeritus astatine the National Technical University of Athens, presented AI Pharos, noting that Greece is 1 of conscionable 7 European countries to articulation the EU’s AI Factories program. “We are trying to align regulatory frameworks with real-world developments,” helium said.

Greece and OpenAI signed a memorandum of understanding successful September to broaden entree to AI tools successful secondary acquisition and beforehand innovation among small businesses. The statement makes Greece 1 of the archetypal countries to instrumentality ChatGPT Edu, a version of ChatGPT tailored for world institutions.

At the aforesaid time, Greece is acceptable to go the archetypal European Union state to enforce nonstop restrictions connected κοινωνικά media use by minors. A caller bill, expected to walk successful Parliament successful the coming weeks, would artifact entree to platforms such arsenic TikTok, Instagram and Facebook for children under 15. Under the proposal, immoderate attempts to entree these platforms from devices used by children would beryllium automatically prevented.

“We person to use technology to enrich students’ skills while respecting each the indispensable information constraints,” Kollias said. “We besides request to spot however AI tin beryllium implemented successful acquisition to train the caller generation.”

Kollias besides pointed to the challenges posed by the accelerated gait of technological improvement and underlined the request to equilibrium innovation with strong regulatory frameworks.

The league was moderated by Kathimerini English Edition’s lawman editor, Nektaria Stamouli.

Kathimerini English Edition Deputy Editor successful Chief Nektaria Stamouli (left) speaks with Marios Dafnomilis (center), advisor connected strategic foresight for the Greek Presidency, and Stefanos Kollias (right), president of National Infrastructures for Research and Technology.

Labor and business

In a sheet titled “The AI Factory arsenic a Business Model,” moderated by Kathimerini English Edition Editor successful Chief Tom Ellis, Christoforos Anagnostopoulos from QuantumBlack, Afrodite Sevasti, main technology serviceman astatine Growthfund, and Michalis Kassimiotis, managing manager for Hewlett Packard successful Greece and Cyprus, debated whether the AI mill is simply a precursor to the emergence of a caller signifier of capitalism, where economies of standard nary longer subordinate to goods and services oregon to code, but to intelligence.

For Anagnostopoulos, a technology chap spouse astatine the consulting limb of McKinsey & Company, “there’s been a surprise development, which is unfastened sourcing. AI could person been a wholly behind-closed-doors idea, blocked down paywalls.”

He besides underscored however SMEs could payment from AI, particularly fixed however the outgo is going down the much radical use it, while noting that infrastructure is indispensable connected the way forward.

Referring to Greece and its request for a high-skilled workforce, Anagnostopoulos said that “the quality to unrecorded successful Greece but service a planetary marketplace will beryllium the biggest origin successful encephalon gain.”

The displacement to AI could dramatically boost nationalist growth, according to Kassimiotis, who besides warned that success would beryllium connected the availability of skilled workers.

“In the Industrial Revolution, standard was based connected producing much efficiently; successful the integer age, it was services. In the AI era, the system of standard is based connected creativity, which tin transform depletion patterns successful a state – oregon adjacent worldwide,” helium said.

He described AI arsenic “a almighty accelerator crossed sectors” that is creating caller forms of competition. But helium besides raised cardinal questions: “How tin these capabilities present existent benefits? And who is ultimately the consumer?”

On policy, Kassimiotis said Europe’s strict regulatory stance connected AI risked holding backmost innovation. “There’s a large debate: innovation archetypal oregon regularisation first. I judge regularisation arsenic it stands does not work – when you restrict innovation, it doesn’t work. Europe cannot beryllium strong lone successful regulating things,” helium said.

Sevasti, who heads the technology conception of Greece’s nationalist concern fund, confirmed that the state is actively “seeking the talent” that will pb the country’s technological advancement efforts.

“Our work is to look astatine the concern facet of however such an endeavor tin payment a society,” she said. “We tin make solutions that person much than 1 applicability. For example, we tin assistance day-to-day activities oregon minimize risks successful agriculture.”

Speaking connected different panel, the vice president of Greece’s National Center for Scientific Research “Demokritos” and technical manager of Pharos AI Factory, Evangelos Karkaletsis, expressed assurance that with the close investments and policies, Greece could go a Southeastern Europe AI hub. “For Greece, technological autonomy is simply a realistic extremity that will alteration its presumption successful the planetary economy,” helium said, linking the inaugural to the European High-Performance Computing program.

European Parliament Member Brando Benifei said the European Commission is under unit to dilute the AI Act, urging afloat enforcement to safeguard antiauthoritarian processes. “We request rules to support democracy, from tackling disinformation to preventing transgression activity,” helium said.

Disinformation was besides a cardinal interest for Martyna Bilandziukiewicz of the European External Action Service, who cited caller attempts to manipulate elections successful Poland and Moldova utilizing AI-generated contented and bots. She called connected journalists and fact-checkers to follow AI tools themselves to antagonistic disinformation.

On the information front, Despina Spanou, cybersecurity advisor astatine the European Commission, noted that 85 percent of ransomware incidents already trust connected AI. Pointing to a caller lawsuit involving Anthropic’s Claude system, she warned, “The astir unsafe risks are the ones we don’t yet know, arsenic we are entering uncharted territory.” Still, she added, AI besides carries “huge potential” for skills improvement and planetary cooperation.

Last but not least, Ann Kristin Glenster, lawman manager astatine the Minderoo Center for Technology and Democracy astatine the University of Cambridge and enforcement manager of the Glenlead Center, and Charalambos Tsekeris, acting seat astatine the Greek National Commission for Bioethics and Technoethics, and main researcher successful integer sociology astatine the Greek National Center for Social Research, joined Kathimerini’s Iliana Magra for a treatment connected “Dangers and Opportunities: ‘Sand boxes and harmless harbors,’ Monopoly hazard and innovation, level economics and the aboriginal of work.”

They explored level economics, labour disruption and what the aboriginal of work means successful an AI-driven economy. “In terms of jobs, AI is hitting the mediate of the workforce. It tin already take connected roles successful accounting oregon law, but it can’t bash plumbing oregon electrical work, and while it can’t yet execute encephalon surgery, 1 time it may,” said Glenster.

“The existent question isn’t lone what jobs we’re losing, but what the jobs that stay will look like. There will beryllium caller opportunities, adjacent jobs maintaining and repairing the technology itself. But it’s hard to foretell which sectors will endure. Manual work and professions rooted successful quality enactment – therapists, teachers, compassionate doctors – will stay essential.”

According to Glenster, “the question present is whether regularisation is the close spot to look for answers, particularly when Europe seems to beryllium falling behind. Humans are not interchangeable with AI, and we request safeguards that support quality autonomy. We indispensable beryllium arsenic bold arsenic the architects of Bretton Woods aft World War II. Ultimately, humans are the astir valuable assets we person – and that’s why I stay optimistic.”

Tsekeris is not acrophobic for the aboriginal of jobs, but “for the archetypal clip successful history, though, there’s a immense spread betwixt the advancement of labour issues and our understanding of them. We request to study these processes cautiously successful bid to modulate globalization. Of course, productivity and ratio substance – but supra all, we indispensable absorption connected creativity and the augmentation of quality thought. Otherwise, we hazard surviving successful a fragmented world. Automation is conscionable 1 broadside of the coin; the different is augmentation.”

“People request much clip to understand and prepare. We should determination beyond fearfulness and take a much pragmatic attack to AI,” helium concluded.