A Brave New World: Overexposure and the Cost of Ambivalence 

By Carmen Gray

If Huxley’s Brave New World demonstrated a society numbed by pleasure and distraction, today’s AI-filled feeds and billboards suggest a certain modern iteration, one in which the limits of our conscience and ethics feel increasingly threatened by convenience.

In the same way we got used to the radio, the TV, and the DVD player, it seems many people have resigned to AI becoming a natural part of life. Suddenly, hearing that a university essay or TV ad was produced by AI is altogether unsurprising and the use of AI in daily life feels unavoidable. Many of us find ourselves guilty of this ambivalence, gladly benefiting from the convenience of AI whilst ignoring the risk. However, this convenience comes at a cost. The dangers of AI are vast, and the media has had no short time reporting on this. However, not all warnings are equal. Articles suggesting that AI will ‘end the human race’ or other sensationalist statements are so vague and exaggerated that they risk being dismissed entirely. 

Sensationalism about AI may have led to its most subtle threat yet- ambivalence.

 This week a new study was released, including over 1,700 participants from across England and Wales. The report (Crest, 2025) examines public attitudes to deepfakes. Deepfakes are a product of generative AI, using an extensive database with images of a particular person to create a lifelike impersonation often indistinguishable from reality. This footage is endless in its bounds; however, most reported on is the use of AI deepfakes in pornography and child exploitation. The Crest report (2025) found that of the respondents, one in four agreed with, or felt neutral about, the legal and moral acceptability of viewing, sharing, creating or selling a sexual or intimate deepfake, regardless of the person’s consent. 

Whilst the devastating impacts of deepfake technology on victims cannot be understated, the point here is not to catalogue its harms. Rather, to discuss why it is that society seems to have grown complacent to the morally reprehensible effects of AI. 

One topic of recent discourse that may explain some of this ambivalence is news fatigue. With constant access to multiple news streams on our mobile, reports have investigated how this new form of news consumption impacts public attitudes to news. A 2025 Al-Jazeera report (Al-Jazeera, 2025) investigating news consumption during the pandemic found that 2/3 of respondents reported being ‘exhausted’ by the volume of news they received. The report noted that this phenomenon may have created scepticism surrounding news and journalism, leading to news avoidance.

How this may apply to AI is versatile. Similarly to news, AI has begun to permeate all forms of social media and platforms, being utilised by a significant portion of advertising, content creators and even news channels themselves. If this weren’t satisfactory alone, articles and news about AI have also had a dominating presence in media, taking up much of the news space with both valid and sensationalised concerns over it’s use and place in society.

This can lead to ‘desensitisation’, that is, an individual’s or society’s reduction in sensitivity to extremes. A 2019 study (Thompson et al., 2019) which documented media consumption over 3 years, concluded that repeated exposure to media led to an eventual ‘numbing’ of those consumers to news previously considered high-interest. It may come as no surprise then, that patterns follow similarly for overexposure to AI. The more it is presented in daily aspects of our lives- advertising, content consumption and so forth, the less shocking it becomes.

To investigate further, I interviewed a student who regularly uses AI to see how this trend of ambivalence may have begun gaining traction across different areas of the UK and Ireland.

The student from Queens University Belfast (who chose to remain anonymous) noted that he tries to use AI ‘sparingly’. Suggesting constraints of AI to university work, he acknowledged the limits of AI in relation to humanities topics. However, he noted that his exposure to AI was ‘exponentially high’- suggesting that he sees AI either in photos or videos ‘every time (I) open my phone’. This particular comment is concerning, given a recent study (Griffith, 2024) suggests that we open our phones 58 times a day. Even if AI exposure is only a handful of this, the significance of the number is distinct. 

This overexposure to AI suggests merit in the concept of news exhaustion, and how it can impact our consumption of media. When asked if he felt overexposed to AI, he responded with certainty that the level of exposure was higher than he would like. Overexposure of AI, he argued, ‘normalises’ its presence in everyday life. Highlighting the presence of AI “art” in the advertising of major businesses, he recounted many developing establishments in Belfast which have resorted to an advertising campaign which relies solely on AI. This large-scale exposure, he argued, legitimises lower expectations on businesses and their creativity. Interestingly, this adds a new level to the permeation of AI. Rather than being confined to our phones; or your personal level of time spent on social media, AI on billboards is an ‘entirely unavoidable’ form of exposure. 

He also noted that the more businesses partake in this practice, the less the public will have the ability to discern between what is real and what is doctored. The images, he emphasised, ‘do not say they are AI; but you can tell’. This ability to tell, however, may actually be in decline as exposure increases. 

Last month, a study from the University of Reading (Gray et al., 2025) noted that without any training, subjects were only able to tell an AI generated face approximately 31% of the time. Of studies in this area, this result is a particularly low percentage. Studies across 2023-2024 suggested that the majority of respondents recognised AI doctored photos when compared with original. This can be seen in a University of Waterloo study (2024) and a 2023 study (Lu et al., 2023) which both note numbers within the range of 60-70% correct classification. Since then, the University of Reading study finding that only 31% of respondents recognised AI indicates a significant drop in the public’s recognition of AI doctored images. 

These trends point towards a significant normalisation of AI in everyday life, with our recognition of AI reducing and its presence ever increasing. Our newfound ambivalence to it, perhaps should be considered unsurprising. This ambivalence, however, remains a threat. The original statistic prompting this question remains a problem; that 1 in 4 are positive or ambivalent about the use of AI deepfakes. 

Taking recent data together, these statistics and interview reflect a broader societal shift toward widespread indifference to the presence of AI. The moral justification of deepfakes exposes how collective societal standards surrounding consent and dignity can be undermined simply on the face of overexposure and normalisation. News fatigue, which emerged with the rise of technology and instant-access internet, was perhaps just the beginning of a society-wide digital saturation. We are less shocked, less concerned, and sadly less surprised, at the widespread injustices which are being reported on. 

As to how these trends will develop, perhaps regulation, education, or public debate will help. Or perhaps we will quietly adjust, normalising what should unsettle us. What seems certain is that this is a test of attention and values rather than technology alone. How we respond to that test determines how deeply AI will reshape our conscience, and the extent to which we allow ourselves to become numb not only to the digital world, but to the news; happy or heartbreaking; happening in our real world every day. 

Bibliography 

Crest (2025). Examining Public Attitudes to Deepfakes. [online] Crest Advisory. Available at: https://www.crestadvisory.com/examining-public-attitudes-to-deepfakes [Accessed 1 Dec. 2025].

Gray, K.L.H., Davis, J.P., Bunce, C., Noyes, E. and Ritchie, K.L. (2025). Training human super-recognizers’ detection and discrimination of AI-generated faces. Royal Society Open Science, 12(11), pp.250921–250921. doi:https://doi.org/10.1098/rsos.250921.

Griffith, N. (2024). 39 Mobile Phone Usage Statistics & Data. [online] SupplyGem. Available at: https://supplygem.com/publications/mobile-phone-usage-statistics/.

Kabashi, O. (2025). News Fatigue and Avoidance: How Media Overload is Reshaping Audience Engagement. [online] Al Jazeera Media Institute. Available at: https://institute.aljazeera.net/en/ajr/article/2826.

Lu, Z., Huang, D., Bai, L., Liu, X., Qu, J. and Ouyang, W. (2023). Seeing is not always believing: A Quantitative Study on Human Perception of AI-Generated Images. [online] arXiv.org. doi:https://doi.org/10.48550/arXiv.2304.13023.

Newman, N. (2024). Overview and key findings of the 2024 Digital News Report | Reuters Institute for the Study of Journalism. [online] reutersinstitute.politics.ox.ac.uk. Available at: https://reutersinstitute.politics.ox.ac.uk/digital-news-report/2024/dnr-executive-summary.

Thompson, R.R., Jones, N.M., Holman, E.A. and Silver, R.C. (2019). Media exposure to mass violence events can fuel a cycle of distress. Science Advances, [online] 5(4), p.eaav3502. doi:https://doi.org/10.1126/sciadv.aav3502.

University of Waterloo (2024). Can You Tell AI-generated People from Real ones? [online] ScienceDaily. Available at: https://www.sciencedaily.com/releases/2024/03/240306003456.htm.