Do We Need Policies to Regulate Modern AIs Right Now?

By Hersh Sanghvi

This is the fourth post in a series about artificial intelligence, along with its uses and social/political implications.

For a long time, Machine Learning (ML) and Artificial Intelligence (AI) have remained firmly in the domain of research applications and science fiction. This is changing thanks to the emergence of a huge variety of AI devices, from smartphones and wearables to robots and autonomous drones. Many companies are working on AI hardware and software as a way to help businesses and organizations create better, smarter solutions, like personal assistants and self-driving cars. As it advances, AI has begun to have an impact on everyday life. We are starting to see more people using AI-powered devices to help their daily lives. The question is: will it help or hurt us?

The previous paragraph, except the first sentence, was written entirely by a special kind of ML model called a “language model”… Despite recent progress, current AI still has many limitations that can make its widespread use dangerous. In light of this, it’s important to understand what these limitations are and ultimately how we as a society should take action. In particular, we’ll look at two exciting, consumer-facing applications of AI: CV and NLP. To get more insight into this topic, I interviewed Dr. CJ Taylor, professor of Computer and Information Science here at UPenn, whose research focuses on computer vision and robotics.

Read more

A Deep Dive into Deepfakes

By Hannah Kolev

This is the third post in a series about artificial intelligence, along with its uses and social/political implications.

An enchanting video of the Spanish artist Salvador Dalí - who passed away in 1989 - remarking on the current weather with visitors of the Dalí Museum. A video of British soccer star David Beckham speaking nine different languages - only one of which he actually speaks - petitioning world leaders to end malaria. A faked pornographic video used to discredit investigative journalist Rana Ayyub for speaking out against a sex abuse case. In each of these scenarios, deepfake videos were used to dupe their audiences. These deepfakes can be used to ignite the imagination, to inspire change, or to intimidate victims. But what are deepfakes and how are they made? What is their purpose and how are they regulated? We will address these questions in the following blog post.

Read more

AI is only as green as the electricity it uses and the problems it solves

by Shannon Wolfman

This is the second post in a series about artificial intelligence, along with its uses and social/political implications.

If you’ve been following the news on climate change and artificial intelligence over the last few years, you might feel conflicted about the potential for AI to help us in combating global warming. For the most part, mainstream and tech publications either exalt AI as a climate savior or decry AI’s ever-increasing carbon footprint. When both issues are discussed in the same article, the focus is on whether AI is doing more harm than good for the environment.

But AI isn’t inherently carbon-intensive; it’s a tool that has tremendous capabilities for mitigating climate damage, and any technology is only as green as its power source. Considering AI’s climate costs and potential climate benefits as distinct issues and understanding the impacts they have on each other can lead to more coherent conversations about the role that this technology can and should play in our climate future.

Read more

Human Cognition in Artificial Intelligence: Bridging the Gap Between Mind and Machine

By Amanda N. Weiss

This is the first post in a series about artificial intelligence, along with its uses and social/political implications.

When you hear the phrase “artificial intelligence,” many different thoughts may come to mind. Perhaps you think of the helpful internet assistants on your phone and computer. Maybe your mind goes to the villainous computers of science fiction stories, seeking revenge on humanity. Whichever contexts your mind wanders into, it is clear that certain types of artificial intelligences have risen in prominence in both in everyday life and in popular media. At its core, artificial intelligence (AI) is the collection of computers or computational programs that are able to exhibit behaviors we would associate with human intelligence. AI algorithms can analyze data and incorporate new information into future decision-making. However, does the ability to learn in this sense truly constitute intelligence? Could advancements in AI eventually lead to sentient, self-aware computers? Or is there something unique to human psychology that cannot be modeled by algorithms and data processing? These are all questions that arise as we continue to use AI technology in our daily lives. In order to gain experienced insight, I had a conversation with Camilo Fosco, a machine learning PhD student at MIT, about artificial intelligence in the context of his research on computer vision and cognition.

Read more

Government and politics in science and technology policy: a necessary relationship that must not be overlooked

By Wisberty J. Gordián Vélez

This is the second post on the big idea of the role of government funding in scientific research.

Knowledge and technologies that we often take for granted, such as the internet, Google search, global positioning system (GPS) devices and magnetic resonance imaging (MRI), are the product of federal investments in research and development (R&D). Federally-supported research promotes innovation: 30% of issued patents rely on this support, which includes government-owned patents, patents citing federal funding, or patents citing other supported patents and research. Scientific and technological innovations account for most of the exponential growth in individual income since 1880. During the COVID-19 pandemic, we have benefited from vaccines developed in part thanks to decades of federally-supported research. The NIH also partnered with Moderna to develop their vaccine, and the Trump administration allocated billions of dollars to develop and manufacture vaccines. The science and technology (S&T) output of the country is directly tied to our individual and collective wellbeing, and the government plays an irreplaceable and necessary role in creating policies that determine what is achieved. In this blog, I explore federal funding of R&D and S&T policy in the U.S., and talk about this with Dr. Kenneth Evans, an S&T policy scholar at Rice University’s Baker Institute for Public Policy and a member of the project staff for a report by the American Academy of Arts and Sciences titled “The Perils of Complacency: America at a Tipping Point in Science & Engineering”.

Read more

More than bleaching: coral cells in hot water

By Lulu Allen-Walker

This is a guest post from contributing author Lulu Allen-Walker. If you are interested in contributing an entry, please contact PSPDG.

The headlines roll in like waves, more every year. Half of the great barrier reef is dead. Reefs are battered by climate change and look “ravaged by war.” But how does ocean warming actually affect how corals function? And can some corals take the heat? I’m on a research team at Penn Biology that’s trying to find answers. Our newest results suggest that heat-stressed corals slow their metabolisms and lose the power to regulate their cellular chemistry – even if they appear healthy at first glance.

Read more

The Role of the United States Government in Funding STEM Research: A Brief History

By Amanda N. Weiss

This is the first post on the big idea of the role of government funding in scientific research.

The United States gained its independence as a country near the start of the first Industrial Revolution. Thus, perhaps unsurprisingly, technological research has been an element of our society for much of its existence. As the country has grown and advanced, it has undergone changes in global involvements and societal priorities, and these have been reflected in the STEM research that the federal government promotes and funds. I recently spoke with Dr. Thomas Cornell, a professor of Science, Technology, and Society at the Rochester Institute of Technology, who offered insight into the ways that government funding of STEM research in the United States has changed over time, especially during periods of upheaval such as the World Wars.

Read more

Political polarization and bias during the COVID-19 pandemic

By Wisberty J. Gordián Vélez

This is our third post on the big idea of confirmation bias.

The United States is experiencing an unprecedented intermingling of crises: job losses not seen since World War II, a pandemic that has killed more than 500,000 Americans, and political division that led to an attack on the Capitol. These crises have heightened our understanding of the role of politics and policy in our lives, as reflected by the record levels of votes cast and turnout in the 2020 election. This historic engagement has been driven in part by political polarization, a phenomenon in which the beliefs of different groups regarding policy, ideology, and political institutions become increasingly oppositional. When driven to extremes, it can impair democracy and the implementation of policies that address society’s problems. Voters are now more likely to dislike the other side and see it as an existential threat to the country at levels exceeding differences in policy opinions. Political scientists have argued that polarization has been fomented by the nationalization of politics and by parties becoming more homogeneous and identifiable with specific policies, social views, race, religion, ideology, and identity. Policymakers also contribute with how they communicate and act in response to a polarized electorate to maintain their power. Every disagreement is a battle to the death between two sides, where cooperation is impossible and no victory is secure. This high-stakes feeling is reflected in the small margins and few districts or states that determine control of Congress and the presidency. As people retreat to their corners, the two parties are differentiated further and polarization is reinforced in a vicious cycle.

Read more

Bias in machine learning models

By Sonia Roberts

This is the second post on the big idea of confirmation bias.

“Confirmation bias” refers to a characteristically human weakness: The tendency to favor information that supports what we already believe. But if we aren’t careful, biases can also pop up in systems that use machine learning tools. Let’s say you want to develop a tool to determine the locations in a city where more crimes are committed so that you can alert the police to send more officers to those locations. The tool is trained using existing police reports. More crimes are reported in the locations where there are more officers, because officers are the ones filing reports. Thus, the tool learns that more crimes are committed in the places that already have a greater density of police officers, and keeps advising the police force to send officers to those locations. There could be other parts of the city with lots of crimes that are not being reported, but the tool will never learn about those crimes -- they simply don’t appear in the dataset. This could create a feedback loop that looks and behaves a lot like confirmation bias does in humans.

Read more

The evolutionary advantages and social pitfalls of confirmation bias

By Amanda N. Weiss

This is the first post on the big idea of confirmation bias.

In a world with an overwhelming abundance of information and opinions, it is no surprise that we do not take it all in and instead must devote our attention to a select subset of information . But, as with every instance of selective information-seeking, we risk our cognitive biases preventing us from forming a well rounded and well reasoned mental model of the world. Confirmation bias, the tendency to seek out information that agrees with our beliefs and reject that which opposes them, is one such element of our psychology. I recently had a discussion with Dr. Tali Sharot, the director of the Affective Brain Lab at the University College London, about her work on the neuroscience of information-seeking and how it relates to confirmation bias. Dr. Sharot is especially interested in figuring out why we engage in behaviors that seem arational (not based on reason), as opposed to those with clear purposes.

Read more

2020 blog format!

For 2020, we here at the PSPDG blog are going to try something a little bit different. Instead of having one-off posts by individual authors, we are going to try posting a series of entries that all touch on the same big idea from the perspective of different fields. Each writer on our team will interview experts from one area and ask them questions about how their research relates to the big theme. We hope you enjoy this in-depth examination of big ideas from multiple perspective.

Where Does Science Fit on Capitol Hill?

As Dr. Lucy Jones remarked in the closing lecture of the AAAS meeting, it is insufficient to bring science to the table, as we must go further to activate it. Activated science is relevant, it is understandable, and it is actionable. Simply explaining scientific information is not always what is needed at a given moment, especially if it doesn’t directly lead to realistic next steps.

Read more

AAAS Annual Meeting 2019: Science Transcending Boundaries

In this 3-part series, I will explore the following main ideas of this conference: the collectivity of science, the intersection of science and policy, and the complexity of science communication. With some personal reflection on my takeaways from the meeting, I will delve into the interface of science and the public, and its implications for society as a whole.

Read more

Event Recap: Milan Yager on Falling Out of Love with Science (video to come!)

by Erin K. Reagan

On November 28th, the Penn Science Policy and Diplomacy Group was thrilled to welcome distinguished speaker Milan Yager, the Executive Director of the American Institute for Medical and Biological Engineering, to the University of Pennsylvania. His topic: Falling Out of Love with Science-- Why Congress Doesn't Fund Medical Innovation. Attendees heard about all the many innovations we use every day which were only made possible through generous federal funding for scientific research, from GPS technology to the iPhone screen. Mr. Yager also discussed the divide between many American voters and the scientists toiling away in their labs, as well as how to bridge the gap between the two seemingly very different worlds. Spoiler alert: the key is seeing the humanity in all people and taking time to understand each person's circumstances before casting aspersions on their opinions of things that might matter greatly to YOU but not to them.

Mr. Yager concluded by issuing a challenge to the audience: become pen pals with your legislators. Email them, write them, go to their town halls and campaign events and bother them about the issues that matter to you, because if you don't, you will quickly learn the truth of what Senator Mike Enzi once famously said: "If you're not at the table, you're on the menu."

After concluding his remarks, Mr. Yager stuck around for a vigorous Q&A session which centered on topics such as the most effective methods for influencing your representatives, how to craft questions in town halls to push politicians to give you real answers, and how to best take advantage of tools and resources offered by institutions like AIMBE and the American Association for the Advancement of Science (AAAS). After the session, a small group of students was able to pick Mr. Yager's brain over lunch for advice and insight on how to make the biggest impact as a student still in training. Everyone enjoyed exchanging some great ideas, as well as sharing some excellent tacos.

Before he hopped back on a train to D.C., we were able to catch Mr. Yager for a quick interview. We'll be putting up the video of our interview soon, so stay tuned!

Podcast - Molly Sheehan: Bioengineer & PA Congressional Candidate




Ian interviews Molly Sheehan, postdoctoral fellow at the University of Pennsylvania's department of bioengineering, and candidate for Pennsylvania's 7th Congressional District.
I chat with Molly about what it's like to run for office while still working in the lab and being a mother, her experiences so far on the campaign, what got her into politics, and her suggestions for college, graduate or medical students - or other postdocs - who are interested in getting politically engaged, but don't know where to start.

Primary elections are on May 15th, 2018
General elections are on November 6th, 2018
If you're interested in volunteering for her, visit: mollysheehan.org/contact/

Opinion: Clearing the Air About Unenforceable Policies


Opinion: Clearing the Air About Unenforceable Policies

 by Kristina Victoreen

Is a policy without enforcement really a policy, or is it just an aspiration? That question has been on my mind lately, in two different contexts, both related to the air we breathe. First, there’s Penn’s new “Tobacco-free Campus” policy. I first noticed the signs in November, when they quietly popped up here and there around campus. As someone who has spent many a lunch hour going from bench to bench all around campus in an often-vain attempt to find a place to sit and eat my lunch without having to breathe second-hand smoke, I was really excited to see those signs. But I confess I was much less excited when I went online and read the actual policy, particularly the section on enforcement. You can read it here.

What it seems to say is that there is no enforcement, and if you have any questions, ask the person you report to or your Dean. In other words, Penn wants you to not smoke but if you do, probably nothing will happen. This idea that the policy won’t be enforced, was confirmed in Rahul Chopra’s December 3rd DP article, in which Frank Leone, Director of the Comprehensive Smoking Treatment Program at Perelman School of Medicine, was quoted as stating that "There's not going to be enforcement or an effort to corral smokers." So Penn’s idea is to try to change the norms, and also provide supports for those who are trying to quit, perhaps partly by removing some of the triggers. For example, the smoking pole outside Van Pelt Library has been removed and replaced with a sign, with the expectation that folks won’t just stand where the pole used to be and drop their cigarette buts on the ground. I’m definitely not an expert on smoking cessation or behavioral economics principles and I know lots of research and some testing went into choosing this approach. Presumably robust baseline data have been collected on smoking behaviors, so that the success of the program can be measured with real outcomes, and I will be very interested to see the results, (and to enjoy a smoke free outdoor lunch when the weather gets warmer.) Certainly it’s no longer unusual to use "nudge" techniques to try to elicit desired behaviour changes, and such policies are popular because they are non-coercive and can be very cost effective. The alternative would be to have the campus police enforce the tobacco policy, and I’m guessing that this may be viewed by the Administration as much more trouble than it's worth, perhaps alienating the people the policy targets, and diverting resources from campus police who have other more pressing concerns. 

In Philadelphia, diesel and other vehicles are subject to several anti-idling laws, enforced (in theory) by different agencies. You can see them all in one place at this helpful site from Pennsylvania Diesel Difference. For example, you can be issued a ticket for $101 by the Philadelphia Parking Authority for excessive idling, and the Department of Health’s Air Management Services can issue a citation to the operator of a heavy duty diesel truck, bus or other vehicle under a separate law, for idling over 2 minutes. There are many exceptions, having to do with things like ambient temperature, (look here for the details) which make the laws incredibly difficult to enforce even if any agency were inclined to enforce them. In addition to Philadelphia’s laws, Pennsylvania has a separate diesel idling law that can be enforced by the State Police. Confused yet? Here’s an experiment to try. Next time you see a PPA agent giving out tickets, try to report an idling vehicle. You might get a quizzical look. I tried this only once, but the PPA officer I asked did not seem to have heard of the anti-idling law. There are a few No Idling signs here and there, but you have to look hard to find them. Thanks to the Clean Air Council, there is a web site where anyone can report an idling vehicle. But it’s doubtful that citations will be issued on the basis of only a citizen complaint, especially without a video to show how long the vehicle idled, and the citizen needs to know the law and be willing to do the reporting.

In short, there are many laws, little enforcement, and no incentive for compliance. So what’s the solution? Should the City be employing nudge methods, and/or trying to change the culture around idling? What would it take to do that? Should the PPA be issuing tickets? I am guessing that a $100 ticket may be seen as a reasonable cost of doing business for the operator of even a small fleet. What about higher fines? According the the New York State web site, there you can be fined up to $18,000 for a first offense with certain idling violations. It seems that steep fines might generate funds to pay for some grants for replacing older engines and doing clean diesel retrofits, but you still need enforcement in order to collect those fines. So at least in the case of vehicle emissions, it appears that policies without enforcement sometimes amount to little more than hope, and as Rudy Giuliani famously said, hope is not a strategy.