Artificial intelligence-generated summaries of scientific papers make complex information more understandable for the public compared with human-written summaries, according to my recent paper published in PNAS Nexus. AI-generated summaries not only improved public comprehension of science but also enhanced how people perceived scientists.
I used a popular large language model, GPT-4 by OpenAI, to create simple summaries of scientific papers; this kind of text is often called a significance statement. The AI-generated summaries used simpler language – they were easier to read according to a readability index and used more common words, like “job” instead of “occupation” – than summaries written by the researchers who had done the work.
In one experiment, I found that readers of the AI-generated statements had a better understanding of the science, and they provided more detailed, accurate summaries of the content than readers of the human-written statements.
I also investigated what effects the simpler summaries might have on people’s perceptions of the scientists who performed the research. In this experiment, participants rated the scientists whose work was described in the simpler texts as more credible and trustworthy than the scientists whose work was described in the more complex texts.
In both experiments, participants did not know who wrote each summary. The simpler texts were always AI-generated, and the complex texts were always human-generated. When I asked participants who they believed wrote each summary, they ironically thought the more complex ones were written by AI and simpler ones were written by humans.
Have you ever read about a scientific discovery and felt like it was written in a foreign language? If you’re like most Americans, new scientific information is probably hard to understand – especially if you try to tackle a science article in a research journal.
In an era where scientific literacy is crucial for informed decision-making, the abilities to communicate and grasp complex ideas are more important than ever. Trust in science has been declining for years, and one contributing factor may be the challenge of understanding scientific jargon.
This research points to a potential solution: using AI to simplify science communication. By making scientific content more approachable, this work demonstrates that AI-generated summaries may help to restore trust in scientists and, in turn, encourage greater public engagement with scientific issues. The question of trust is particularly important, as people often rely on science in their daily lives, from eating habits to medical choices.
What still isn’t known
As AI continues to evolve, its role in science communication may expand, especially if using generative AI becomes more commonplace or sanctioned by journals. Indeed, the academic publishing field is still establishing norms regarding the use of AI. By simplifying scientific writing, AI could contribute to more engagement with complex issues.
While the benefits of AI-generated science communication are perhaps clear, ethical considerations must also be considered. There is some risk that relying on AI to simplify scientific content may remove nuance, potentially leading to misunderstandings or oversimplifications. There’s always the chance of errors, too, if no one pays close attention.
Additionally, transparency is critical. Readers should be informed when AI is used to generate summaries to avoid potential biases.
Simple science descriptions are preferable to and more beneficial than complex ones, and AI tools can help. But scientists could also achieve the same goals by working harder to minimize jargon and communicate clearly – no AI necessary.
Antibiotic resistance is a growing public health problem around the world. When bacteria like E. coli no longer respond to antibiotics, infections become harder to treat.
To develop new antibiotics, researchers typically identify the genes that make bacteria resistant. Through laboratory experiments, they observe how bacteria respond to different antibiotics and look for mutations in the genetic makeup of resistant strains that allow them to survive.
While effective, this method can be time-consuming and may not always capture the full picture of how bacteria become resistant. For example, changes in how genes work that don’t involve mutations can still influence resistance. Bacteria can also exchange resistance genes between each other, which may not be detected if only focusing on mutations within a single strain.
My colleagues and Ideveloped a new approach to identify E. coli resistance genes by computer modeling, allowing us to design new compounds that can block these genes and make existing treatments more effective.
Identifying resistance
To predict which genes contribute to resistance, we analyzed the genomes of various E. coli strains to identify genetic patterns and markers associated with resistance. We then used machine learning algorithms trained on existing data to highlight novel genes or mutations shared across resistant strains that might contribute to resistance.
After identifying resistance genes, we designed inhibitors that specifically target and block the proteins these genes produce. By analyzing the structure of the proteins these genes code for, we were able to optimize our inhibitors to strongly bind to these specific proteins.
To reduce the likelihood that bacteria would evolve resistance to these inhibitors, we targeted regions of their genome that code for proteins critical to their survival. By interfering with how bacteria carry out important functions, it makes it more difficult for them to develop mechanisms to compensate. We also prioritized compounds that work differently from existing antibiotics to minimize cross-resistance.
Finally, we tested how effectively our inhibitors could overcome antibiotic resistance in E. coli. We used computer simulations to assess how strongly a number of inhibitors bind to target proteins over time. One inhibitor called hesperidin was able to strongly bind to the three genes in E. coli involved in resistance that we identified, suggesting it may be able to help combat antibiotic-resistant strains.
By targeting the specific genes responsible for resistance to existing drugs, our approach could lead to treatments for challenging bacterial infections that are not only more effective but also less likely to contribute to further resistance. It can also help researchers keep up with bacterial threats as they evolve.
Some microbes can transfer resistance to other microbes.
Our predictive approach could be adapted to other bacterial strains, allowing for more personalized treatment strategies. In the future, doctors could potentially tailor antibiotic treatments based on the specific genetic makeup of the bacteria causing the infection, potentially leading to better outcomes.
As antibiotic resistance continues to rise globally, our findings may provide a crucial tool in the fight against this threat. Further development is needed before our methods can be used in the clinic. But by staying ahead of bacterial evolution, targeted inhibitors could help preserve the efficacy of existing antibiotics and reduce the spread of resistant strains.
On Sept. 20, 2024, a newspaper in Montana reported an issue with ballots provided to overseas voters registered in the state: Kamala Harris was not on the ballot. Election officials were able to quickly remedy the problem but not before accusations began to spread online, primarily among Democrats, that the Republican secretary of state had purposefully left Harris off the ballot.
This false rumor emerged from a common pattern: Some people view evidence such as good-faith errors in election administration through a mindset of elections being untrustworthy or “rigged,” leading them to misinterpret that evidence.
As the U.S. approaches another high-stakes and contentious election, concerns about the pervasive spread of falsehoods about election integrity are again front of mind. Some election experts worry that false claims may be mobilized – as they were in 2020 – into efforts to contest the election through tactics such as lawsuits, protests, disruptions to vote-counting and pressure on election officials to not certify the election.
Our team at the University of Washington has studied online rumors and misinformation for more than a decade. Since 2020, we have focused on rapid analysis of falsehoods about U.S. election administration, from sincere confusion about when and where to vote to intentional efforts to sow distrust in the process. Our motivations are to help quickly identify emerging rumors about election administration and analyze the dynamics of how these rumors take shape and spread online.
Through the course of this research we have learned that despite all the discussion about misinformation being a problem of bad facts, most misleading election rumors stem not from false or manipulated evidence but from misinterpretations and mischaracterizations. In other words, the problem is not just about bad facts but also faulty frames, or the mental structures people rely on to interpret those facts.
Misinformation may not be the best label for addressing the problem – it’s more an issue of how people make sense of the world, how that sensemaking process is shaped by social, political and informational dynamics, and how it begets rumors that can lead people to a false understanding of events.
Rumors – not misinformation
There is a long history of research on rumors going back to World War II and earlier. From this perspective, rumors are unverified stories, spreading through informal channels that serve informational, psychological and social purposes. We are applying this knowledge to the study of online falsehoods.
Though many rumors are false, some turn out to be true or partially true. Even when false, rumors can contain useful indications of real confusions or fears within a community.
Rumors can be seen as a natural byproduct of collective sensemaking – that is, efforts by groups of well-meaning people to make sense of uncertain and ambiguous information during dynamic events. But rumors can also emerge from propaganda and disinformation campaigns that lead people to misinterpret or mischaracterize their own and others’ experiences.
University of Washington’s Kate Starbird explains rumors as collective sensemaking.
Evidence, frames and (mis)interpretations
Prior research describes collective sensemaking as a process of interactions between evidence and frames. Evidence includes the things people see, read and hear in the world. Frames are mental schema that shape how people interpret that evidence.
The relationship between evidence and frames flows in two directions. When people encounter novel events or new evidence, they try to select the best frame from their mental filing cabinets. The selected frame then determines what evidence they focus on and what evidence they exclude in their interpretations. This evidence-frame view of collective sensemaking can help researchers understand rumors and disinformation.
Everyone has their own ways of interpreting events based on their unique experiences. But your frames are not yours alone. Frames are shaped, sometimes intentionally, by information from media, political leaders, communities, colleagues, friends, neighbors and family. Framing – the process of using, building, reinforcing, adapting, challenging and updating frames – can be a deliberate strategy of political communication.
Frames play a role in generating rumors, shaping how people interpret emerging events and novel evidence. False rumors occur when sensemaking goes awry, often due to people focusing on the wrong piece of evidence or applying the wrong frame. And disinformation, from this perspective, is the intentional manipulation of the sensemaking process, either by introducing false evidence or distorting the frames through which people interpret that evidence.
In 2020, we saw these dynamics at work in a rumor about Sharpie pens in Arizona. In the lead-up to the election, President Donald Trump and his allies repeatedly alleged that the election would be rigged – setting a powerful frame for his followers. When voters noted that the Sharpie pens provided by election officials were bleeding through their ballots, many interpreted their experiences through the frame of a “rigged election” and became concerned that their ballots would not be counted.
A Maricopa County, Arizona, election worker counts ballots in the 2020 election as false rumors that Sharpie pens were ruining ballots spread online. AP Photo/Matt York
Some people shared those experiences online, where they were soon amplified and given meaning by others, including online influencers. Concerns and suspicions grew. Soon, members of Trump’s family were repeating false claims that the bleed-through was systematically disenfranchising Republican voters. The effect was circular and mutually reinforcing. The strategic frame inspired misinterpretations of evidence – real bleed-through falsely seen as affecting ballot counting – that were shared and amplified, strengthening the frame.
Social media sensemaking
Collective sensemaking is increasingly taking place online, where it is profoundly shaped by social media platforms, from features such as repost and like buttons to algorithmic recommendations to the connections between accounts.
Not so long ago, many people hoped that the internet would democratize information flows by removing the historical gatekeepers of information and disrupting their ability to set the agenda – and the frames – of conversation. But the gatekeepers have not been erased; they have been replaced. A group of newsbrokering influencers have taken their place, in part by gaming the ways online systems manipulate attention.
Many of these influencers work by systematically seeking out and amplifying content that aligns with prevailing political frames set by elites in politics and media. This gives creators the incentive to produce content that resonates with those frames, because that content tends to be rewarded with attention, the primary commodity of social media.
These dynamics were at work in February 2024, when an aspiring creator produced a man-on-the-street video interviewing migrants to the U.S. that was selectively edited and captioned to falsely claim to show undocumented migrants planning to vote illegally in U.S. elections. This resonated with two prominent frames: the same rigged-election frame from 2020 and another that framed immigration as harmful to the U.S.
The video was shared across multiple platforms and exploded in views after being amplified by a series of accounts with large followings on X, formerly Twitter. X CEO Elon Musk commented with an exclamation point on one post with the embedded video. The creator soon found himself on Fox News. He currently has hundreds of thousands of followers on TikTok and Instagram and continues to produce similar content.
Interactions between influencers and online audiences result in content that fits strategic frames. Emerging events provide new evidence that people can twist to fit prevailing frames, both intentionally and unintentionally. Rumors are the byproducts of this process, and online attention dynamics fuel their spread.
Collective sensemaking and election 2024
Heading into the 2024 election, false and misleading claims about election integrity remain widespread. Our team has tracked more than 100 distinct rumors since the beginning of September. The machinery for quickly converting perceived evidence from elections into widely shared rumors and conspiracy theories is increasingly well oiled.
Experts discuss election integrity and efforts to undermine voter confidence.
One concerning development is an increase in so-called election integrity organizations that seek to recruit volunteers who share the rigged-election frame. The groups aim to provide volunteers with tools to streamline the collection and amplification of evidence to support the rigged-election frame.
One worry is that these volunteers may misinterpret what they see and hear on Election Day, generating additional rumors and false claims about election integrity that reinforce that increasingly distorted frame. Another is that these false claims will feed lawsuits and other attempts to contest election results.
However, we hope that by shedding light on some of these dynamics, we can help researchers, journalists, election officials and other decision-makers better diagnose and respond to rumors about election integrity in this cycle. Most importantly, we believe that this collective sensemaking lens can help us all to both empathize with well-meaning people who get caught up in sharing false rumors and see how propagandists manipulate these processes for their gain.
It’s Halloween. You’ve just finished trick-or-treating and it’s time to assess the haul. You likely have a favorite, whether it’s chocolate bars, peanut butter cups, those gummy clusters with Nerds on them or something else.
For some people, including me, one piece stands out – the Snickers bar, especially if it’s full-size. The combination of nougat, caramel and peanuts coated in milk chocolate makes Snickers a popular candy treat.
As a food engineer studying candy and ice cream at the University of Wisconsin-Madison, I now look at candy in a whole different way than I did as a kid. Back then, it was all about shoveling it in as fast as I could.
Now, as a scientist who has made a career studying and writing booksabout confections, I have a very different take on candy. I have no trouble sacrificing a piece for the microscope or the texture analyzer to better understand how all the components add up. I don’t work for, own stock in, or receivefunding from Mars Wrigley, the company that makes Snickers bars. But in my work, I do study the different components that make up lots of popular candy bars. Snickers has many of the most common elements you’ll find in your Halloween candy.
Let’s look at the elements of a Snickers bar as an example of candy science. As with almost everything, once you get into it, each component is more complex than you might think.
Snickers bars contain a layer of nougat, a layer of caramel mixed with peanuts and a chocolate coating.
Let’s start with the nougat. The nougat in a Snickers bar is a slightly aerated candy with small sugar crystals distributed throughout.
One of the ingredients in the nougat is egg white, a protein that helps stabilize the air bubbles that provide a light texture. Often, nougats like this are made by whipping sugar and egg whites together. The egg whites coat the air bubbles created during whipping, which gives the nougat its aerated texture.
A boiled sugar syrup is then slowly mixed into the egg white sugar mixture, after which a melted fat is added. Since fat can cause air bubbles to collapse, this step has to be done last and very carefully.
The final ingredient added before cooling is powdered sugar to provide seeds for the sugar crystallization in the batch. The presence of small sugar crystals makes the nougat “short” – pull it apart between your fingers and it breaks cleanly with no stretch.
Chewy caramel
On top of the nougat layer is a band of chewy caramel. The chewiness of the caramel contrasts the nougat’s light, airy texture, which provides contrast to each bite.
Caramel stands out from other candies as it contains a dairy ingredient, such as cream or evaporated milk. During cooking, the milk proteins react with some of the sugars in a complex series of reactions called Maillard browning, which imparts the brown color and caramelly flavor.
Maillard browning starts with proteins and certain sugars. The end products of these reactions include melanoidins, which are brown coloring compounds, and a variety of flavors. The specific flavor molecules depend on the starting materials and the conditions, such as temperature and water content.
Commercial caramel, like that in the Snickers bar, is cooked up to about 240-245 degrees Fahrenheit (115-118 degrees Celsius), to control the water content. Cook to too high a temperature and the caramel gets too hard, but if the cook temperature is too low, the caramel will flow right off the nougat. In a Snickers bar, the caramel needs to be slightly chewy so the peanuts stick to it.
Chocolate coating
To make chocolate, raw cocoa beans are harvested from cacao pods and then fermented for several days. After the fermented beans are dried, they are roasted to develop the chocolate flavor. As in caramel, the Maillard browning reaction is an important contributor to the flavor of chocolate.
The milk chocolate coating on the Snickers bar happens through a process called enrobing. The naked bar, arranged on a wire mesh conveyor, passes through a curtain of tempered liquid chocolate, covering all sides with a thin layer. Tempering the chocolate coating makes it glossy and gives it a well-defined snap.
The flow of the tempered chocolate needs to be controlled precisely to give a coating of the desired thickness without leading to tails at the bottom of the candy bar.
The Snickers bar
When done right, the result is a delicious Snickers bar, a popular Halloween – or anytime – candy.
With about 15 million bars made each day, getting every detail just right requires a lot of scientific understanding and engineering precision.
