🧠 52% Fear AI, Do You?

Subscribe | Contact | Meet The Team

In today’s Sunday Special:

Read Time: 5 minutes

🤥PUBLIC DISTRUST PERMEATES

The U.S., despite concerns about the strength of its democracy, takes public opinion seriously. Though embarrassingly low approval ratings may indicate otherwise, politicians discuss issues some voters care about. For better or for worse, they regurgitate narrow, stingy concerns during TV appearances and Congressional testimonies to energize their base. Reactionary politics, coupled with widespread, consistent coverage of important, niche issues like transgender athletes, critical race theory in K–12 schools, and corporate activism, demonstrates that negative emotion-evoking issues receive more attention from the political class.

Increasingly, AI has become a flashpoint in public discourse. Last month, Pew Research found that 52% of Americans feel more concerned than excited about artificial intelligence in their daily lives. In contrast, just 10% reported more excitement than concern, and the remaining 38% had mixed feelings. If you’re reading this, you likely fall into the latter two cohorts. Are Americans misinformed about AI’s cost-benefit tradeoff? Or is the crowd wiser than it seems at first glance?

👨‍⚖️ARE CONCERNS JUSTIFIED?

Less than 1 in 3 Americans can correctly identify six simple uses of AI, so they may fear what they don’t understand. But these not-so-wise crowds might have sound intuition. Their sentiments match those of hundreds of AI experts and leaders who signed an open letter in March, calling upon labs and companies to pause experiments more powerful than GPT-4. Separately, over 45% of Americans hold a high-risk occupation, meaning more than half of their daily tasks are currently susceptible to automation. These include the following job categories:

Other employment-related concerns include:

Unsurprisingly, job-related AI concerns are top of mind for respondents.

Less charitably, negative or unrealistic AI portrayals in popular media may influence the concerned cohort. In The Terminator, robots try to destroy the human species, and Ex Machina anthropomorphizes AI, giving it human-like qualities like self-awareness and deception, which are unlikely to manifest in the foreseeable future. Though film depictions of AI are, on balance, positive, negativity is more memorable, and it stacks on top of employment-related concerns. It’s also plausible that the concerned cohort isn’t aware of the immense AI-driven benefits they’ve reaped. In an Edelman AI Center of Expertise survey, most respondents didn’t believe that tech companies used AI in Facebook photo tagging, Google searches, or Amazon recommendations. Finally, the 52% may immediately recall recent experiences with dodgy automated customer support or robocalls, conflating the quality of audio-based AI applications with more advanced prediction models like recommendation algorithms or chatbots.

Excited respondents, on the other hand, likely associate gains in convenience, productivity, safety, healthcare, and information accessibility with AI. College-educated workers, overrepresented among the 10%, are most exposed to AI in their jobs, and their tasks are most likely to be complemented by AI. Within each line of work, an AI copilot or machine-learning wizard would improve the average productivity per worker, thus reducing the absolute number of jobs. A percentage of professionals in select fields, hypothetically the bottom quartile of radiologists, data analysts, software engineers, financial analysts, or lawyers by performance, might be replaced with a novel class of jobs—machine-learning practitioners and prompt engineers—requiring subject matter expertise and data science competency. Thus, the excitement of professionals may not be justified.

🤖IT MAY NOT MATTER

Regardless of individual dispositions, we wonder if public concerns preceded the mass adoption of revolutionary technology in the past. During the PC revolution, many experts feared job loss, social unrest, and even the end of humanity. Cyberphobia struck the public, and several publications devoted coverage to appeasing their concerns. When the internet held similar consciousness in the public’s mind as AI does today, just 6% expressed dislike for “computers and technology.” Intuitively, Americans on the cutting edge of innovation have always been out of touch with the broader public, but since their inventions provided immense value, it didn’t matter. In this sense, AI adoption might follow a similar upward trajectory, despite lingering concerns.

Conversely, AI’s trifecta of capability—speed, scale, and disruptive capacity—may inflict harm widely and rapidly. The internet could send messages faster than humans, but deep fake images and audio can trick sophisticated users. More troubling is the potential of autonomous software agents that complete tasks as programmed or the perpetual jostling match between distributors and detectors of misinformation, malicious software, and deepfakes.

Diving deeper into Americans’ concerns, they are more pronounced in AI applications where slight underperformance causes physical harm (e.g., self-driving) or culturally unacceptable outcomes(e.g., bias in hiring). Although this is a rational inclination, the average American’s risk perception is not. Folks tend to overestimate the risk of death by flight, sharks, lightning, and a variety of other calamities covered in popular media. Moreover, people are more averse to losses than attracted to gains of similar size. Perception of and aversion to AI-related risks will be no different. If history is any guide, consumers will vote with their wallets, or more aptly, their attention, barring a cataclysmic event.

📒FINAL NOTE

If you found this useful, follow us on Twitter or provide honest feedback below. It helps us improve our content.

🚨Subscribe to our newsletter for free and receive these powerful Notion templates: