Medical AI could be ‘dangerous’ for poorer nations, WHO warns

Curated on Posted on by Stefaan Verhulst

Article by David Adam: “The introduction of health-care technologies based on artificial intelligence (AI) could be “dangerous” for people in lower-income countries, the World Health Organization (WHO) has warned.

The organization, which today issued a report describing new guidelines on large multi-modal models (LMMs), says it is essential that uses of the developing technology are not shaped only by technology companies and those in wealthy countries. If models aren’t trained on data from people in under-resourced places, those populations might be poorly served by the algorithms, the agency says.

“The very last thing that we want to see happen as part of this leap forward with technology is the propagation or amplification of inequities and biases in the social fabric of countries around the world,” Alain Labrique, the WHO’s director for digital health and innovation, said at a media briefing today.

The WHO issued its first guidelines on AI in health care in 2021. But the organization was prompted to update them less than three years later by the rise in the power and availability of LMMs. Also called generative AI, these models, including the one that powers the popular ChatGPT chatbot, process and produce text, videos and images…(More)”.

2024 Edelman Trust Barometer

Curated on Posted on by Stefaan Verhulst

Edelman: “The 2024 Edelman Trust Barometer reveals a new paradox at the heart of society. Rapid innovation offers the promise of a new era of prosperity, but instead risks exacerbating trust issues, leading to further societal instability and political polarization.

Innovation is accelerating – in regenerative agriculture, messenger RNA, renewable energy, and most of all in artificial intelligence. But society’s ability to process and accept rapid change is under pressure, with skepticism about science’s relationship with Government and the perception that the benefits skew towards the wealthy.

There is one issue on which the world stands united: innovation is being poorly managed – defined by lagging government regulation, uncertain impacts, lack of transparency, and an assumption that science is immutable. Our respondents cite this as a problem by nearly a two to one margin across most developed and developing countries, plus all age groups, income levels, educational levels, and genders. There is consensus among those who say innovation is poorly managed that society is changing too quickly and not in ways that benefit “people like me” (69%).

Many are concerned that Science is losing its independence: to Government, to the political process, and to the wealthy. In the U.S., two thirds assert that science is too politicized. For the first time in China, we see a contrast to their high trust in government: Three-quarters of respondents believe that Government and organizations that fund research have too much influence on science. There is concern about excessive influence of the elites, with 82% of those who say innovation is managed poorly believing that the system is biased in favor of the rich – this is 30 percentage points higher than those who feel innovation is managed well…(More)”.

Integrating Participatory Budgeting and Institutionalized Citizens’ Assemblies: A Community-Driven Perspective

Curated on Posted on by Stefaan Verhulst

Article by Nick Vlahos: “There is a growing excitement in the democracy field about the potential of citizen’s assemblies (CAs), a practice that brings together groups of residents selected by lottery to deliberate on public policy issues. There is longitudinal evidence to suggest that deliberative mini-publics such as those who meet in CAs can be transformative when it comes to adding more nuance to public opinion on complex and potentially polarizing issues.

But there are two common critiques of CAs. The first is that they are not connected to centers of power (with very few notable exceptions) and don’t have authority to make binding decisions. The second is that they are often disconnected from the broader public, and indeed often claim to be making their own, new “publics” instead of engaging with existing ones.

In this article I propose that proponents of CAs could benefit from the thirty-year history of another democratic innovation—participatory budgeting (PB). There are nearly 12,000 recorded instances of PB to draw learnings from. I see value in both innovations (and have advocated and written about both) and would be interested to see some sort of experimentation that combines PB and CAs, from a decentralized, bottom-up, community-driven approach.

We can and should think about grassroots ways to scale and connect people across geography using combinations of democratic innovations, which along the way builds up (local) civic infrastructure by drawing from existing civic capital (resident-led groups, non-profits, service providers, social movements/mobilization etc.)…(More)”.

People Have a Right to Climate Data

Curated on Posted on by Stefaan Verhulst

Article by Justin S. Mankin: “As a climate scientist documenting the multi-trillion-dollar price tag of the climate disasters shocking economies and destroying lives, I sometimes field requests from strategic consultants, financial investment analysts and reinsurers looking for climate data, analysis and computer code.

Often, they want to chat about my findings or have me draw out the implications for their businesses, like the time a risk analyst from BlackRock, the world’s largest asset manager, asked me to help with research on what the current El Niño, a cyclical climate pattern, means for financial markets.

These requests make sense: People and companies want to adapt to the climate risks they face from global warming. But these inquiries are also part of the wider commodification of climate science. Venture capitalists are injecting hundreds of millions of dollars into climate intelligence as they build out a rapidly growing business of climate analytics — the data, risk models, tailored analyses and insights people and institutions need to understand and respond to climate risks.

I point companies to our freely available data and code at the Dartmouth Climate Modeling and Impacts Group, which I run, but turn down additional requests for customized assessments. I regard climate information as a public good and fear contributing to a world in which information about the unfolding risks of droughts, floods, wildfires, extreme heat and rising seas are hidden behind paywalls. People and companies who can afford private risk assessments will rent, buy and establish homes and businesses in safer places than the billions of others who can’t, compounding disadvantage and leaving the most vulnerable among us exposed.

Despite this, global consultants, climate and agricultural technology start-ups, insurance companies and major financial firms are all racing to meet the ballooning demand for information about climate dangers and how to prepare for them. While a lot of this information is public, it is often voluminous, technical and not particularly useful for people trying to evaluate their personal exposure. Private risk assessments fill that gap — but at a premium. The climate risk analytics market is expected to grow to more than $4 billion globally by 2027.

I don’t mean to suggest that the private sector should not be involved in furnishing climate information. That’s not realistic. But I worry that an overreliance on the private sector to provide climate adaptation information will hollow out publicly provided climate risk science, and that means we all will pay: the well-off with money, the poor with lives…(More)”.

Representative Bodies in the Age of AI

Curated on Posted on by Stefaan Verhulst

Report by POPVOX: “The report tracks current developments in the U.S. Congress and internationally, while assessing the prospects for future innovations. The report also serves as a primer for those in Congress on AI technologies and methods in an effort to promote responsible use and adoption. POPVOX endorses a considered, step-wise strategy for AI experimentation, underscoring the importance of capacity building, data stewardship, ethical frameworks, and insights gleaned from global precedents of AI in parliamentary functions. This ensures AI solutions are crafted with human discernment and supervision at their core.

Legislatures worldwide are progressively embracing AI tools such as machine learning, natural language processing, and computer vision to refine the precision, efficiency, and, to a small extent, the participatory aspects of their operations. The advent of generative AI platforms, such as ChatGPT, which excel in interpreting and organizing textual data, marks a transformative shift for the legislative process, inherently a task of converting rules into language.

While nations such as Brazil, India, Italy, and Estonia lead with applications ranging from the transcription and translation of parliamentary proceedings to enhanced bill drafting and sophisticated legislative record searches, the U.S. Congress is prudently venturing into the realm of Generative AI. The House and Senate have initiated AI working groups and secured licenses for platforms like ChatGPT. They have also issued guidance on responsible use…(More)”.

Experts in Government

Curated on Posted on by Stefaan Verhulst

Book by Donald F. Kettl: “From Caligula and the time of ancient Rome to the present, governments have relied on experts to manage public programs. But with that expertise has come power, and that power has long proven difficult to hold accountable. The tension between experts in the bureaucracy and the policy goals of elected officials, however, remains a point of often bitter tension. President Donald Trump labeled these experts as a ‘deep state’ seeking to resist the policies he believed he was elected to pursue—and he developed a policy scheme to make it far easier to fire experts he deemed insufficiently loyal. The age-old battles between expertise and accountability have come to a sharp point, and resolving these tensions requires a fresh look at the rule of law to shape the role of experts in governance…(More)”.

Facts over fiction: Why we must protect evidence-based knowledge if we value democracy

Curated on Posted on by Stefaan Verhulst

Article by Ben Feringa and Paul Nurse: “Central to human progress are three interconnected pillars. The first is pursuit of knowledge, a major component of which is the expansion of the frontiers of learning and understanding – something often achieved through science, driven by the innate curiosity of scientists.

The second pillar of progress is the need for stable democracies where people and ideas can mix freely. It is this free exchange of diverse perspectives that fuels the democratic process, ensuring policies are shaped by a multitude of voices and evidence, leading to informed decision-making that benefits all of society.

Such freedom of speech and expression also serves as the bedrock for scientific inquiry, allowing researchers to challenge prevailing notions without fear, fostering discovery, applications and innovation.

The third pillar is a fact-based worldview. While political parties might disagree on policy, for democracy to work well all of them should support and protect a perspective that is grounded in reliable facts, which are needed to generate reliable policies that can drive human progress….(More)”.

Testing the Assumptions of the Data Revolution

Curated on Posted on by Stefaan Verhulst

Report by TRENDS: “Ten years have passed since the release of A World that Counts and the formal adoption of the Sustainable Development Goals (SDGs). This seems an appropriate time for national governments and the global data community to reflect on where progress has been made so far. 

This report supports this objective in three ways: it evaluates the assumptions that underpin A World that Counts’ core hypothesis that the data revolution would lead to better outcomes across the 17 SDGs, it summarizes where and how we have made progress, and it identifies knowledge gaps related to each assumption. These knowledge gaps will serve as the foundation for the next phase of the SDSN TReNDS research program, guiding our exploration of emerging data-driven paradigms and their implications for the SDGs. By analyzing these assumptions, we can consider how SDSN TReNDs and other development actors might adapt their activities to a new set of circumstances in the final six years of the SDG commitments.

Given that the 2030 Agenda established a 15-year timeframe for SDG attainment, it is to be expected that some of A World that Counts’ key assumptions would fall short or require recalibration along the way. Unforeseen events such as the COVID-19 pandemic would inevitably shift global attention and priorities away from the targets set out in the SDG framework, at least temporarily…(More)”.

Tackling Today’s Data Dichotomy: Unveiling the Paradox of Abundant Supply and Restricted Access in the Quest for Social Equity

Curated on Posted on by Stefaan Verhulst

Article by Stefaan Verhulst: “…One of the ironies of this moment, however, is that an era of unprecedented supply is simultaneously an era of constricted access to data. Much of the data we generate is privately “owned,” hidden away in private or public sector silos, or otherwise inaccessible to those who are most likely to benefit from it or generate valuable insights. These restrictions on access are grafted onto existing socioeconomic inequalities, driven by broader patterns of exclusion and marginalization, and also exacerbating them. Critically, restricted or unequal access to data does not only harm individuals: it causes untold public harm by limiting the potential of data to address social ills. It also limits attempts to improve the output of AI both in terms of bias and trustworthiness.

In this paper, we outline two potential approaches that could help address—or at least mitigate—the harms: social licensing and a greater role for data stewards. While not comprehensive solutions, we believe that these represent two of the most promising avenues to introduce greater efficiencies into how data is used (and reused), and thus lead to more targeted, responsive, and responsible policymaking…(page 22-25)”.

What does it mean to trust a technology?

Curated on Posted on by Stefaan Verhulst

Article by Jack Stilgoe: “A survey published in October 2023 revealed what seemed to be a paradox. Over the past decade, self-driving vehicles have improved immeasurably, but public trust in the technology is low and falling. Only 37% of Americans said they would be comfortable riding in a self- driving vehicle, down from 39% in 2022 and 41% in 2021. Those that have used the technology express more enthusiasm, but the rest have seemingly had their confidence shaken by the failure of the technology to live up to its hype.

Purveyors and regulators of any new technology are likely to worry about public trust. In the short term, they worry that people won’t want to make use of new innovations. But they also worry that a public backlash might jeopardize not just a single company but a whole area of technological innovation. Excitement about artificial intelligence (AI) has been accompanied by a concern about the need to “build trust” in the technology. Trust—letting one’s guard down despite incomplete information—is vital, but innovators must not take it for granted. Nor can it be circumvented through clever engineering. When cryptocurrency enthusiasts call their technology “trustless” because they think it solves age-old problems of banking (an unavoidably imperfect social institution), we should at least view them with skepticism.

For those concerned about public trust and new technologies, social science has some important lessons. The first is that people trust people, not things. When we board an airplane or agree to get vaccinated, we are placing our trust not in these objects but in the institutions that govern them. We trust that professionals are well-trained; we trust that regulators have assessed the risks; we trust that, if something goes wrong, someone will be held accountable, harms will be compensated, and mistakes will be rectified. Societies can no longer rely on the face-to-face interactions that once allowed individuals to do business. So it is more important than ever that faceless institutions are designed and continuously monitored to realize the benefits of new technologies while mitigating the risks….(More)”.