91��ɫ

By 'Damola Adediji, Shadi Nasseri, Dominic Rochon, and Xiang Zhang

For , Associate Professor and 91��ɫ Research Chair in Artificial Intelligence, Data Governance, and the Law, understanding that moment of hesitation has become a central scholarly pursuit. Following the publication of his recent book, , Penney sat down with IP Osgoode researchers to discuss why people often silence themselves in the face of surveillance, uncertainty, and power, and why prevailing legal theories fail to adequately explain this phenomenon.

The origins of Penney's work can be traced back to the summer of 2013, when Edward Snowden's disclosures exposed the unprecedented scale of global mass surveillance. The revelations transformed public debates about privacy and state power, but they also raised a deeper question: what happens to human behaviour when people know they might be watched?

For Penney, the answer lies in the phenomenon known as “chilling effects.” Typically, in legal scholarship, chilling affects are said to occur when individuals refrained from speaking, searching, associating, or participating in public life because they perceive a risk of adverse consequences. Yet Penney argues that the conventional understanding of chilling effects is too narrow.

For decades courts and scholars have largely viewed chilling effects through an individualistic lens. Under this traditional account, individuals are assumed to be rational actors who engage in cost benefit calculations: they encounter a vague law, fear illegal sanction, and choose to self-censor. But what if the person who abandoned that Google search was not calculating legal risks at all?

We like to think of ourselves as rational and independent individuals who think for ourselves and do not ‘follow the herd,’ Penney explained. Research in social psychology, however, paints a different picture. Drawing on Solomon Asch’s famous conformity experiments and a broader body of behavioural scholarship, Penney argues that chilling effects are better understood as a form of conformity and compliance. Human beings are deeply social creatures whose behaviour is shaped by norms, expectations, and perceptions of authority. Often, we modify our conduct not because we have been directly threatened, but because we sense that certain actions may expose us to scrutiny or disapproval. In this sense, the hesitation before a sensitive search query is not merely an individual decision. It reflects a broader social and psychological process.

“We are, at a very fundamental level, wired to be chilled,” Penney observed. Throughout human history, conformity in response to threats promoted safety and survival. In the digital age, many of those same instincts are activated by surveillance technologies, data collection practices, and online monitoring. To understand when chilling effects are most likely to occur, Penney identifies four key factors.

The first is observation. People are more likely to modify their behaviour when they know or suspect that they're being watched. Digital surveillance systems amplify this effect by making observation pervasive, persistent, and often invisible.

The second factor is uncertainty. The less certain individuals are about how rules will be interpreted or enforced, the more likely they are to conform. Ambiguous laws, opaque algorithms, and unclear policies can all generate chilling effects.

Third is personalization. Chilling affects intensify when individuals feel specifically targeted or singled out. Personalized surveillance, threat, or interventions are often more powerful than generalized monitoring.

Finally, there is power and authority. People are more likely to alter their behaviour when observation or intervention originates from institutions that possess significant authority, whether governments, employers, platforms, or other powerful actors.

Together, these factors provide a framework for understanding why seemingly voluntary acts of self-censorship emerge across a wide range of contexts. Importantly, Penney’s theory is supported by empirical evidence. Following the Snowden revelations, he conducted a large-scale study examining traffic to Wikipedia articles related to privacy-sensitive topics, including “terrorism” and “political violence.” The findings were striking. Views of those articles dropped by approximately 25 percent after the disclosures and remained depressed for more than a year. The significance of this study was not that people stopped engaging in unlawful conduct. Rather that millions of ordinary individuals appeared less willing to seek information about lawful but potentially sensitive subjects.

Returning to our opening example, the concern is not whether someone searched for “terrorism.” The concern is that countless people may have decided not to search at all. This distinction becomes particularly important when considering whether chilling effects are always harmful. Penney argues that some forms of chilling are socially desirable. We generally want laws to discourage harassment, threats, discrimination, and other harmful conduct. The challenge lies in distinguishing between chilling harmful behaviour and chilling democratic participation.

One example emerged from Penney’s research on anti-harassment protections. Conventional legal theory might predict that such laws would suppress expression. Yet his work found that informing women about these protections often increased their willingness to participate online. Rather than chilling speech, the laws reduced the chilling effects imposed by private abusers. In other words, affective regulation can sometimes create the conditions necessary for greater participation and expression.

These insights have significant implications for contemporary debates about artificial intelligence and technology regulation. Penney noted that powerful technology companies frequently argue that privacy laws, copyright protections, or AI regulations will “chill innovation.” He remains skeptical of such claims and argues that large technology firms possess substantial resources, legal teams, and market power. They are unlikely to stop innovating because of regulation. The more pressing concern, Penney argues, is whether unregulated technologies chill the rights and freedoms of the public. That concern becomes even more urgent as societies move toward what Penney describes as “superveillance”: systems that combine artificial intelligence, facial recognition, big data analytics, and automated enforcement. Such systems bring together all four chilling factors simultaneously. They enable constant observation, generate uncertainty through opaque algorithmic decision making, personalize interventions, and operate through powerful institutions. The result is not simply surveillance, but an environment that encourages conformity and predictability. For Penney, addressing these risks requires more than individual privacy protections. It demands structural reforms, including greater transparency, stronger oversight of AI systems, and restrictions on technologies that pose unacceptable threats to fundamental freedoms.

---