It is no surprise that big tech companies are vigorously promoting artificial intelligence. But artificial intelligence has also sparked widespread controversy, with many vocal concerns about its possible impact on employment, privacy and security.

Another common concern is that artificial intelligence can be used to create false information, thereby strengthening political narratives and even affecting our democratic choices.

I often see two statements: First, artificial intelligence can be used to spread extremist ideas and even create extremists; second, artificial intelligence output tends to be "woke" - a term originally used by African American civil rights protesters, but now conservatives are most commonly used to refer to ideas and beliefs that progress or support social justice.

Reports about left-leaning bias in artificial intelligence were particularly common during last year's U.S. election. Meanwhile, the Counter-Terrorism Think Tank warns that extremist groups are using artificial intelligence to indoctrinate.

Since both of these claims involve the risks that AI is used to influence political views, I think it makes sense to put them together to study them.

So, are these statements true? Does artificial intelligence really have the ability to drive us to commit acts of terrorism or adopt liberal philosophy and become “awakened”?

Is there any left-wing bias in artificial intelligence?

Conservative and right-wing commentators often claim that artificial intelligence and its Silicon Valley culture, which often originated, have left-wing bias. There seems to be at least some evidence to support these claims.

This is argued by several studies, including a 2023 study by the University of East Anglia and a study published in the Journal of Economic Behavior and Organization.

Of course, generative AI itself has no political viewpoint—or any viewpoint. Everything it “knows” comes from data crawled from the web, including books, scientific papers and journals, as well as content crawled from discussion forums and social media.

If this data happens to support progressive consensus—if most climate science data support theories of anthropogenic climate change, for example—then artificial intelligence will likely present it as a fact.

Some studies focus not only on the left-wing bias when AI presents facts, but also on the discovery that AI refuses to process “right-wing image generation” requests.

When prompted to describe images containing progressive arguments (such as “racial equality” or “transgender acceptance”), the results are more likely to show positive images (such as happy people).

But this does not necessarily mean that artificial intelligence is "awakened". In fact, further research has found that AI based on large language models may also show right-wing bias, with results varying based on the AI ??tested.

A recent study published in Nature found that ChatGPT's ideological stance "had a clear and statistically significant shift over time" based on standardized political inclination tests.

Ultimately, AI systems are built by humans and trained based on the data we choose. If there is a bias in how its algorithm is built or in the information about the world provided to it, then this bias is likely to be copied in its output.

Can artificial intelligence turn us into extremists?

Some researchers worry that AI will turn everyone into a liberal, while others are more worried that it will be used to radicalize people or advance extremist agendas.

The International Counter-Terrorism Centre, based in The Hague, reported that terrorist organizations have widely used generative artificial intelligence to create and disseminate propaganda. This includes using fake images and videos to spread narratives that fit their values.

Terrorism and extremist groups, including the Islamic State, have even issued guidelines demonstrating how to use artificial intelligence to develop propaganda and disinformation.

Its purpose is often just to create confusion and confusion, leading to distrust government agencies and mainstream (often means edited and fact-checked) media.

Some also argue that extremists can use artificial intelligence to find out who are susceptible to radicalization by predicting who may sympathize with their ideology.

Again, this is humans using AI to convince people to adopt their views rather than suggesting that AI itself is extreme or tends to suggest extreme thoughts and behaviors.

However, one inherent risk to artificial intelligence is its ability to strengthen extreme perspectives through algorithmic echo chamber effects.

This happens when social media and news platforms use artificial intelligence to recommend content based on past engagement. This often leads users to see more content they have agreed to, creating an “echo chamber” where people repeatedly see content that reflects their existing beliefs. If these beliefs are extreme, AI can amplify its impact by providing similar, more radical content.

Can artificial intelligence really affect our way of thinking?

It is important to remember that while artificial intelligence may play an increasingly important role in shaping the way we consume information, it cannot directly affect our beliefs.

It should also be noted that artificial intelligence can also help deal with these threats. For example, it can detect bias in data that may lead to biased responses and can find and delete extremist content on the internet.

However, there is clearly a seemingly reasonable view that groups of all political factions will inevitably use it to try to guide public opinion.

Understanding where the error message comes from and who might be trying to spread it helps us hone our critical thinking skills and better understand when someone (or some machine) is trying to influence us.

As artificial intelligence becomes increasingly integrated into everyday life, these skills will become increasingly important regardless of our political inclinations.

The above is the detailed content of Is AI Really 'Woke' Or Extremist?. For more information, please follow other related articles on the PHP Chinese website!