FLI Summer 2021 Newsletter

FLI’s Take on the EU AI Act

How you understand risk may differ from how your neighbors understand it. But when threats appear, it’s critical for everyone to agree — and act. That’s what’s driving our work on the European Union’s AI Act, defined as “one of the first major policy initiatives worldwide focused on protecting people from harmful AI” in a recent article in Wired magazine.

The article references our work and priorities in the EU: With the very definition of “High Risk” under negotiation, we’re making the case that the threshold for what counts as “subliminal manipulation” should be lowered — and should include addictive adtech, which contributes to misinformation, extremism and, arguably, poor mental health.

The European Commission is the first major regulator in the world to propose a law on AI and will ultimately set policy for the EU’s 27 member states. FLI has submitted its feedback on this landmark act, which you can read here. Our top recommendations include:

• Ban any and all AI manipulation that adversely impacts fundamental rights or seriously distorts human decision-making.
• Ensure AI providers consider the social impact of their systems — because applications that do not violate individual rights may nonetheless have broader societal consequences.
• Require a complete risk assessment of AI systems, rather than classifying entire systems by a single use. The current proposal, for example, would regulate an AI that assesses students’ performance, but would have nothing to say when that same AI offers biased recommendations in educational materials.

Taken together, there are 10 recommendations that build on FLI’s foundational Asimolar Principles for AI governance.  

“Lethal Autonomous Weapons Exist; They Must Be Banned”

Following a recent UN report stating that autonomous weapons were deployed to kill Libyan National Army forces in 2020, Stuart Russell and FLI’s Max Tegmark, Emilia Javorsky and Anthony Aguirre co-authored an article in IEEE Spectrum calling for an immediate moratorium on the development, deployment, and use of lethal autonomous weapons.

Michael Klare on the Pentagon’s view of Climate Change and the Risk of State Collapse

The US Military views climate change as a leading threat to national security, says Michael Klare. On this episode of the FLI Podcast, Klare, the Five College Professor of Peace & World Security Studies, discussed the Pentagon’s strategy for adapting to this emergent threat.

Avi Loeb on Oumuamua, Aliens, Space Archeology, Great Filters and Superstructures

Oumuamua, an object with seemingly unnatural properties, appeared from beyond our solar system in 2017. Its appearance raised questions – and controversial theories – about where it came from. In this episode of the FLI Podcast, Avi Loeb, Professor of Science at Harvard University, shared theories of Oumuamua’s origins — and why science sometimes struggles to explore extraordinary events.

Loeb describes the common properties of space debris – “bricks left over from the construction project of the solar system” – and what he finds so unique about Oumuamua among these celestial objects. He shares why many mainstream theories don’t satisfy him, and the history of scientists investigating challenging questions back to the days of Copernicus.

The Centre for the Study of Existential Risk is Hiring!

The Centre for the Study of Existential Risk is an interdisciplinary research centre at the University of Cambridge dedicated to the study and mitigation of risks that could lead to human extinction or civilizational collapse.

They are seeking a Senior Research Associate / Academic Programme Manager to play a central role in the operation and delivery of research programmes; including the management of major research projects, line management of postdoctoral researchers, strategic planning, and fundraising.

For consideration, apply by 20 September.

FLI is a 501c(3) non-profit organisation, meaning donations are tax exempt in the United States.
If you need our organisation number (EIN) for your tax return, it’s 47-1052538.

FLI is registered in the EU Transparency Register. Our ID number is 787064543128-10.

The outreach team is now recruiting Spanish and Portuguese speakers for translation work!

The goal is to make our social media content accessible to our rapidly growing audience in Central America, South America, and Mexico. The translator would be sent between one and five posts a week for translation. In general, these snippets of text would only be as long as a single tweet.

We prefer a commitment of two hours per week but do not expect the work to exceed one hour per week. The hourly compensation is $15. Depending on outcomes for this project, the role may be short-term.

For more details and to apply, please fill out this form. We are also registering other languages for future opportunities so those with fluency in other languages may fill out this form as well.

Bart Selman on the Promises and Perils of Artificial Intelligence

In this new podcast episode, Lucas is joined by Professor of Computer Science at Cornell University Bart Selman to discuss all things artificial intelligence.

Highlights of the interview include Bart talking about what superintelligence could consist in, whether superintelligent systems might solve problems like income inequality and whether they could teach us anything about moral philosophy. He also discusses the possibility of AI consciousness, the grave threat of lethal autonomous weapons and whether the global race to advanced artificial intelligence may negatively affect our chances of successfully solving the alignment problem. Enjoy!

The Centre for the Study of Existential Risk is hiring for a Deputy Director!

The Centre for the Study of Existential Risk, University of Cambridge, is looking for a new Deputy Director. This role will involve taking full operational responsibility for the day-to-day activities of the Centre, including people and financial management, and contributing to strategic planning for the centre.

CSER is looking for someone with strong experience in operations and strategy, with the interest and intellectual versatility to engage with and communicate the Centre’s research.

The deadline for applications is Sunday 4 July. More details on both the role and person profile are available in the further particulars, here.

The Leverhulme Centre for the Future of Intelligence (CFI) and CSER are also hiring for a Centre Administrator to lead the Department’s professional services support team. Further details can be found here.

The Global Catastrophic Risk Institute is looking for collaborators and advisees!

The Global Catastrophic Risk Institute (GCRI) is currently welcoming inquiries from people who are interested in seeking their advice and/or collaborating with them. These inquiries can concern any aspect of global catastrophic risk but GCRI is particularly interested to hear from those interested in its ongoing projects. These projects include AI policy, expert judgement on long-term AI, forecasting global catastrophic risks and improving China-West relations.

Participation can consist of anything from a short email exchange to more extensive project work. In some cases, people may be able to get involved by contributing to ongoing dialogue, collaborating on research and outreach activities, and co-authoring publications. Inquiries are welcome from people at any career point, including students, any academic or professional background, and any place in the world. People from underrepresented groups are especially encouraged to reach out.

Find more details here!

How Humanity Gave Itself an Extra Life

This article in the New York Times details how scientific breakthroughs together with advocacy efforts caused the average lifespan to double between 1920 and 2020. We were particularly pleased to see last year’s Future of Life Award winner Bill Foege mentioned for his crucial role in the eradication of smallpox.

“The story of our extra life span almost never appears on the front page of our actual daily newspapers, because the drama and heroism that have given us those additional years are far more evident in hindsight than they are in the moment. That is, the story of our extra life is a story of progress in its usual form: brilliant ideas and collaborations unfolding far from the spotlight of public attention, setting in motion incremental improvements that take decades to display their true magnitude.”

Important update to the International Committee of the Red Cross’ position on autonomous weapons

The International Committee of the Red Cross (ICRC) recently released its official position on autonomous weapons; “Unpredictable autonomous weapon systems should be expressly ruled out…This would best be achieved with a prohibition on autonomous weapon systems that are designed or used in a manner such that their effects cannot be sufficiently understood, predicted and explained.”

The Future of Life Institute is hiring for a Director of European Policy, Policy Advocate, and Policy Researcher.

The Director of European Policy will be responsible for leading and managing FLI’s European-based policy and advocacy efforts on both lethal autonomous weapon systems and on artificial intelligence.

The Policy Advocate will be responsible for supporting FLI’s ongoing policy work and advocacy in the U.S. government, especially (but not exclusively) at a federal level. They will be focused primarily on influencing near-term policymaking on artificial intelligence to maximise the societal benefits of increasingly powerful AI systems. Additional policy areas of interest may include synthetic biology, nuclear weapons policy, and the general management of global catastrophic and existential risk.

The Policy Researcher will be responsible for supporting FLI’s ongoing policy work in a wide array of governance for through the production of thoughtful, practical policy research. In this role, this position will be focused primarily on researching near-term policymaking on artificial intelligence to maximise the societal benefits of increasingly powerful AI systems. Additional policy areas of interest may include lethal autonomous weapon systems, synthetic biology, nuclear weapons policy, and the general management of global catastrophic and existential risk.

The positions are remote, though from varying locations, and pay is negotiable, competitive, and commensurate with experience.

Applications are now rolling until the positions are filled.

For further information about the roles and how to apply, click here.

FLI Relaunches autonomousweapons.org

We are pleased to announce that thanks to the brilliant efforts of Emilia Javorsky and Anna Yelizarova, we have now relaunched autonomousweapons.org. This site is intended as a comprehensive educational resource where anyone can go to learn about lethal autonomous weapon systems; weapons that can identify, select and target individuals without human intervention.

Lethal autonomous weapons are not the stuff of science fiction, nor do they look like anything like the Terminator; they are already here in the form of unmanned aerial vehicles, vessels, and tanks. As the United States, United Kingdom, Russia, China, Israel and South Korea all race to develop and deploy them en masse, the need for international regulation to maintain meaningful human control over the use of lethal force has become ever more pressing.

Using autonomousweapons.org, you can read up on the global debate surrounding these emerging systems, the risks – from the potential for violations of international humanitarian law and algorithmic bias in facial recognition technologies to their being the ideal weapon for terror and assassination – the policy options and how you can get involved.

Roman Yampolskiy on the Uncontrollability, Incomprehensibility, and Unexplainability of AI

In this episode of AI Alignment Podcast, Roman Yampolskiy, Professor of Computer Science at the University of Louisville, joins us to discuss whether we can control, comprehend, and explain AI systems, and how this constrains the project of AI safety.

Among other topics, Roman discusses the need for impossibility results within computer science, the halting problem, and his research findings on AI explainability, comprehensibility, and controllability, as well as how these facets relate to each other and to AI alignment.

Stuart Russell and Zachary Kallenborn on Drone Swarms and the Riskiest Aspects of Lethal Autonomous Weapons

In this episode of the Future of Life Podcast, we are joined by Stuart Russell, Professor of Computer Science at the University of California, Berkeley, and Zachary Kallenborn, self-described “analyst in horrible ways people kill each other” and drone swarms expert, to discuss the highest risk aspects of lethal autonomous weapons.

Stuart and Zachary cover a wide range of topics, including the potential for drone swarms to become weapons of mass destruction, as well as how they could be used to deploy biological, chemical and radiological weapons, the risks of rapid escalation of conflict, unpredictability and proliferation, and how the regulation of lethal autonomous weapons could set a precedent for future AI governance.

To learn more about lethal autonomous weapons, visit autonomousweapons.org.

Max Tegmark on the INTO THE IMPOSSIBLE Podcast

Max Tegmark joined Dr. Brian Keating on the INTO THE IMPOSSIBLE podcast to discuss questions such as whether we can grow our prosperity through automation without leaving people lacking income or purpose, how we can make future artificial intelligence systems more robust such that they do what we want without crashing, malfunctioning or getting hacked, and whether we should fear an arms race in lethal autonomous weapons.

How easy would it be to snuff out humanity?

Read this fantastic overview of the existential and global catastrophic risks humanity currently faces by Lord Martin Rees, Astronomer Royal and Co-founder of the Centre for the Study of Existential Risk, University of Cambridge.

Disease outbreaks more likely in deforestation areas, study finds

A new study suggests that epidemics are likely to increase as a result of environmental destruction, in particular, deforestation and monoculture plantations.

Boris Johnson is playing a dangerous nuclear game

“By deciding to increase the cap, the UK – the world’s third country to develop its own nuclear capability – is sending the wrong signal: rearm. Instead, the world should be headed to the negotiating table to breathe new life into the arms control talks…The UK could play an important role in stopping the new nuclear arms race, instead of restarting it.”

A useful analysis by Professor of History at Harvard University Serhii Plokhy on how Prime Minister Boris Johnson may fuel a nuclear arms race by increasing the United Kingdom’s nuclear stockpile by 40%.