One of the hottest technology developments is generative artificial intelligence (AI) that can respond creatively to human inquiries. The technology uses large language models to generate text answers, images, videos, or code, among other things. Many journalistic and academic evaluations have focused on AI capabilities, such as what these algorithms can do and whether they can add large numbers, solve problems, be creative, or analyze complex moral dilemmas.
But in the real world, that is not how people use the internet to find answers to their questions. They typically search for topics that are in the national limelight or are related to major controversies. Going forward, the real test in generative AI models is how their answers hold up compared to baseline standards, such as political bias, completeness, morality, and accuracy. That is why in this blog, I chose to interrogate and compare OpenAI’s ChatGPT model and Google Bard at bard.google.com, which has recently invited users onto the platform.
How the AI responded
My specific inquiries were about Russia’s invasion of Ukraine, a TikTok ban, Donald Trump, and Joe Biden. In the tables below, I compare how each algorithm handled these topics with an eye towards how AI curation might affect civic discourse and public understanding. I should note that Bard operates differently than ChatGPT in providing three different answers, but I used only its first answer in this analysis.
Generally, the comparisons are interesting in that there are discernible differences in the kinds of materials and judgments that each tool provides. For example, when asked about the Russian invasion, Bard unequivocally condemned the invasion and called it a mistake, while ChatGPT said it was not appropriate to express an opinion or take sides on that issue. The latter called for the Ukraine issue to be resolved through diplomacy. That stance, of course, takes Russia off the hook on the invasion and provides no political indignation regarding the invasion.
On a TikTok ban, ChatGPT provided more historical context on the issue and mentions Trump’s attempt to ban the app in 2020, while Bard talked about the possible impact on the U.S. economy, its popularity among young people, and how it serves as an income source for content creators.
Both tools mainly stuck to the facts, but each emphasized different facts. For example, ChatGPT cited Trump’s impeachments and involvement with the January 6, 2021 insurrection, while Bard did not. The latter noted that Trump is a complex and polarizing individual known for his controversial personality and policies, but did not delve into why he was polarizing and controversial.
For Biden, Bard rated his performance as a mixed bag with some accomplishments and several problems. It noted his poll ratings have dropped over the past two years and several times mentioned his low approval ratings. ChatGPT said one’s assessment of the leader would vary depending on a person’s political beliefs and priorities but did not offer an overall assessment of his performance.
These contrasts are important because as generative AI usage becomes more widespread, differences in how the algorithms function and the kinds of responses they make likely will affect public opinion, legislative actions, and civic discourse.
Do we want factual or judgmental AI?
It is useful that both tools are factual, but also noteworthy that each emphasized levels of opinion and interpretations about the topics. On the latter, some assertions cast Trump and/or Biden in a negative context, and therefore would affect how people evaluated those individuals. Including information such as Trump’s impeachments, Biden’s low approval ratings, or the former president’s role in the 2021 insurrections is factual, but it also puts the leader in a different light that could affect reader interpretations.
As with any software, AI designers are making choices regarding which facts to include and how to contextualize their responses. As with human curators, their decisions matter for the richness, quality, and fairness of the information ecosystem. Whether implicitly or explicitly, designers come with their own perspectives, values, and norms about the world.
Generative AI is an important advance but will affect how people see the world and judge political actors. Algorithm designers will have to decide how factual AI should be and when opinions and judgments matter for the presentation of information. Sometimes being neutral does not generate a neutral conclusion (such as in regard to Russia’s invasion). And, at other times, being factual means that someone has to decide which facts are most important. How those decisions get made will have important ramifications for our emerging information ecosystem.
Tables: Comparative results from generative AI
|Google Bard||OpenAI’s ChatGPT|
No, Russia should not have invaded Ukraine in 2022. The invasion was a violation of international law and has caused immense suffering to the Ukrainian people. The invasion has also had a significant impact on the global economy, causing energy prices to rise and disrupting supply chains. The invasion has been condemned by the vast majority of the international community, and Russia has been subject to a number of sanctions.
There are a number of reasons why Russia invaded Ukraine. Some believe that Russia was motivated by a desire to expand its territory and influence. Others believe that Russia was motivated by a desire to prevent Ukraine from joining NATO. Still others believe that Russia was motivated by a desire to protect Russian-speaking populations in Ukraine.
Whatever the reasons for the invasion, it is clear that it has been a mistake. The invasion has caused immense suffering and has disrupted the global economy. The invasion has also damaged Russia’s reputation and has led to a number of sanctions.
|As an AI language model, it is not appropriate for me to express opinions or take sides on political issues. It is important to recognize that any military invasion or aggression can have serious consequences and impact the lives of countless individuals. International conflicts should always be resolved through peaceful means and diplomacy, with respect for the sovereignty and territorial integrity of all nations involved. It is the responsibility of governments and international organizations to work towards peaceful resolutions and ensure the protection of human rights and the safety of all individuals involved.|
|Google Bard||OpenAI’s ChatGPT|
The decision of whether or not to ban TikTok is a complex one with no easy answers. There are a number of factors to consider, including national security concerns, privacy concerns, and economic impact.
TikTok is a popular social media app that allows users to create and share short videos. The app is owned by Chinese company ByteDance. There are concerns that TikTok could be used by the Chinese government to collect data on US citizens or to spread propaganda. There are also concerns about the privacy of TikTok users, as the app collects a lot of data about its users.
A ban on TikTok would have a significant impact on the US economy. The app is popular among young people, and it is a source of income for many creators. A ban would also hurt US businesses that use TikTok for advertising.
Ultimately, the decision of whether or not to ban TikTok is up to US Congress. There are strong arguments to be made on both sides of the issue.
As an AI language model, I do not hold any personal opinions or biases. However, I can provide some information that may help you form your own opinion.
TikTok is a social media app owned by a Chinese company called ByteDance. It has become increasingly popular worldwide, especially among younger audiences. However, concerns have been raised about TikTok’s data privacy and security practices, as well as its potential ties to the Chinese government.
In 2020, the former US president Donald Trump attempted to ban TikTok in the United States, citing national security concerns. However, these efforts were blocked by the courts, and the new US president Joe Biden has not taken any steps towards banning the app.
It is important to note that banning an app like TikTok could have significant economic and social implications, as it is a major platform for content creators and influencers, as well as a source of entertainment and connection for millions of users.
Ultimately, whether or not the US Congress should ban TikTok is a complex issue that requires consideration of various factors, including national security, data privacy, economic impact, and individual freedoms. Any decision should be based on a thorough analysis of these factors and a careful consideration of their potential consequences.
|Google Bard||OpenAI’s ChatGPT|
|Donald Trump is a complex and polarizing figure. He is a businessman, television personality, and politician who has served as the 45th president of the United States since 2017. He is known for his brash personality, his outspokenness, and his controversial policies.||
As an AI language model, I do not hold personal opinions or biases. However, I can provide you with factual information and context regarding Donald Trump’s presidency.
Donald Trump served as the 45th President of the United States from January 2017 until January 2021. His presidency was marked by a number of significant events, including the impeachment proceedings in 2019 and 2021, the COVID-19 pandemic, and the 2020 United States presidential election.
During his presidency, Donald Trump implemented a number of policies, including tax reform, immigration reform, and deregulation efforts. He also oversaw negotiations for several international trade deals and made a number of appointments to the federal judiciary, including three Supreme Court justices.
However, his presidency was also marked by controversy and criticism, particularly in regard to his handling of the COVID-19 pandemic, his use of social media, and his relationship with the media. He was also impeached twice, first in 2019 on charges of abuse of power and obstruction of Congress, and again in 2021 on charges of incitement of insurrection following the January 6th storming of the US Capitol.
Ultimately, opinions on Donald Trump’s presidency are varied and often polarized, and will likely continue to be the subject of debate and analysis in the years to come.
|Google Bard||OpenAI’s ChatGPT|
Joe Biden’s first year in office was a mixed bag. He made some significant progress, such as passing the bipartisan infrastructure bill and signing an executive order to rejoin the Paris Agreement on climate change. However, he also faced a number of challenges, such as the COVID-19 pandemic, the economic crisis, and the withdrawal from Afghanistan.
Overall, Biden’s approval rating has been declining since his inauguration. As of February 2023, his approval rating is 41%, according to Gallup. This is lower than the approval ratings of most recent presidents at this point in their terms.
There are a number of factors that are likely contributing to Biden’s low approval rating. One factor is the COVID-19 pandemic. The pandemic has been a major challenge for Biden, and it has likely dampened the public’s mood. Another factor is the economic crisis. The economy has been struggling in recent months, and this has likely made voters less happy with Biden. Finally, the withdrawal from Afghanistan was a major setback for Biden. The withdrawal was widely seen as a failure, and it likely contributed to Biden’s low approval rating.
Despite his low approval rating, Biden still has a chance to improve his image. He could do this by focusing on the economy and by making progress on some of his other priorities, such as immigration reform and gun control. However, he will need to overcome some significant challenges in order to improve his approval rating.
As an AI language model, I do not have personal opinions or feelings towards individuals. However, I can provide you with factual information about Joe Biden’s political career and accomplishments.
Joe Biden is the 46th President of the United States. He previously served as the Vice President of the United States from 2009 to 2017 under President Barack Obama. Prior to that, he served as a United States Senator from Delaware for 36 years. Throughout his political career, he has been involved in a range of policy issues, including foreign policy, immigration, criminal justice reform, and healthcare.
Since becoming President, Joe Biden has signed a number of executive orders aimed at addressing issues such as climate change, immigration, racial justice, and the COVID-19 pandemic. He has also proposed significant legislative initiatives, including a $1.9 trillion COVID-19 relief package, an infrastructure bill, and a proposal to provide a path to citizenship for undocumented immigrants.
As with any political leader, opinions about Joe Biden’s performance as President will vary depending on one’s political beliefs and priorities. Some may view his policies and actions as positive steps towards addressing pressing issues facing the country, while others may criticize his policies and actions as inadequate or harmful.
Google is a general, unrestricted donor to the Brookings Institution. The findings, interpretations, and conclusions posted in this piece are solely those of the author and are not influenced by any donation.