After a long day you open Paramount+, scroll past the recommendations and select a show that feels almost predestined for your choice. Eventually, Paramount+ will recommend future content, curated for you based on your watch history. The platform seems to “get” you, what you like, what you skip, even what you want to watch next. While it’s convenient for viewers to click, pause and replay their favorite scenes, most users are unaware that their data is being fed into a system designed to track, and monetize user behavior. AI has fundamentally changed how Paramount+ operates, transforming it from a TV content distributor to a new media entity.
Paramount+, the latest video streaming service from Paramount Global, acts as a traditional media conglomerate comprising a movie studio, one of the major TV studios, and several TV networks, hosting content ranging from sports to entertainment. However, the company has been “tech-forward” since 2023 and uses AI to support content localization, to generate recommendations, to target ads, and to deliver a super-personalized experience.
Most platforms now are relying on machine learning models to maximize revenue, engagement and retention by processing user behaviour data. Creative professionals are being put to work generating data to help improve these systems through data collection and algorithmic optimization.
The Recommendation Engine Knows You Better Than You Think
At the most visible level, AI powers the streaming services recommendation engine. Machine learning algorithms are applied to the clickstream (i.e. what users have watched or not watched, what they’ve searched for, how much engagement has occurred on certain content items etc.) to serve the user the best possible recommendations. These recommendations should get progressively better as the user continues to interact with the platform. While recommendation systems have become a major area of study within streaming media research, much of this scholarship has focused on Netflix’s recommender system, which Pajkovic describes as having received “the most scholarly attention” among algorithmic applications in the film and television industry and was responsible for approximately 80% of total hours streamed on the platform as of 2015. In contrast, platforms such as Paramount+ have received comparatively less academic attention, despite employing similar data-driven personalization strategies.
The Hidden Power of Metadata
One of the less visible but most powerful components of this system is metadata. Using generative AI, Paramount+ automatically produces descriptions, genre tags, and even “micro-genres” for its library of over 50,000 videos.
This metadata does more than organize content. It actively shapes what users see. By categorizing and ranking media in specific ways, the platform influences discovery, deciding which shows are surfaced and which remain buried.
This process turns viewing into a curated experience, where algorithms guide users through content in ways that feel organic but are highly structured. As Paramount describes it, the system functions like a “video DJ,” assembling personalized streams based on predicted preferences. Thus, resulting in a feedback loop: users interact with content, the system learns from those interactions, and future recommendations become even more targeted. By continuously adapting to viewer behaviour, the platform encourages longer viewing sessions and sustained engagement, increasing the likelihood that users remain active on the service.
When Your Data Turns to Profit
So far, we’ve looked at how personalization can keep audiences hooked. More consequentially, it also creates new opportunities for generating revenue. One of the most significant ways this occurs is through the collection and use of audience data for advertising purposes.
Paramount is arming businesses with more tools to reach their audiences on its properties. Paramount Ads Manager, which uses AI to create, target and launch campaigns, allows businesses to reach audiences across the entire Paramount portfolio of properties and platforms. That includes Paramount+, Pluto TV and the CBS broadcast network, as well as several television studios.
By utilizing generative AI technology, the system can generate video ads in minutes from a given company’s website or social media content and air within 24 hours. TV ads can air in as little as days, starting at $500 per listing.
We’ve previously reported on how data from users’ behaviour on Paramount+ is used to suggest more suitable content. Now, it appears this information is also worth a pretty penny to advertisers, as the streaming service increases in value as users become more knowledgeable to potential advertisers.
Generative AI tools, including Adobe Firefly, were used to generate unique images based on user prompts for their campaign. The tool generated content within a matter of seconds and the campaign had a huge level of engagement from audiences. Rather than simply promoting content, these campaigns function as another mechanism for data collection. By tracking which AI-generated advertisements receive the most engagement, Paramount can gain deeper insights into audience preferences and optimize future campaigns accordingly. In this way, AI not only personalizes marketing but also expands the platform’s ability to extract value from user behaviour.
Paramounts Take on AI Replacing Creatives, Reshaping Business
Publicly, Paramount presents a more restrained narrative around artificial intelligence. Executives emphasize that AI is primarily a “support tool” useful for tasks like script analysis or marketing, but not a replacement for human creativity. Paramount’s CTO, Phil Wiser, has stated that AI-generated scripts tend to be “very boring,” and humans will continue to be mainly responsible for producing films for the foreseeable future.
While AI may not be fully replacing creative labor, it is being deeply embedded into the platform’s infrastructure, especially in areas tied to data, marketing, and monetization.
This demonstrates how streaming services have become increasingly datafied, relying on the continuous collection and analysis of audience behaviour. In relation to platformisation, cultural producers and audiences are becoming more dependent on digital infrastructures that generate value through data extraction and algorithmic recommendation. Rather than simply recommending content, AI helps Paramount+ keep users engaged for longer periods of time. The more users interact with the platform, the more behavioural data Paramount+ collects, which can then be used to refine recommendations, target marketing, and support business decisions. This reflects what platform scholars describe as datafication: the transformation of user activity into quantifiable data that can be leveraged for profit.
Paramount has described AI as central to improving search, recommendation, and discovery systems. At the corporate level, EVP and CIO Lakshman Nathan, has framed AI as a way to “turbocharge” content growth and engagement.
The Streaming Wars and the Fight for Attention
This shift is not happening alone. Paramount+ is operating in an increasingly competitive and unstable media landscape.
According to industry analysis, Paramount Global generates roughly $28 billion in annual revenue from a diverse portfolio of movie production and television networks. The bulk of that revenue still comes from old-school channels like advertising, with a sizable portion coming from Paramount’s Direct-to-Consumer division, which includes Paramount+ and Pluto TV, which accounted for approximately 26% of the company’s consolidated revenue in 2024. While the company has seen revenue spike during the digital revolution, its streaming service hasn’t quite turned a profit yet.
Artificial Intelligence (AI) is, perhaps, the biggest new competitor in the fight for human attention. As generative models of AI become more cost-effective, larger players like Netflix, Amazon and Apple can create more content than ever before in attempts to garner screen time. This new development will enhance personalization by enabling experiences and recommendations tailored to individual viewers. Yet, it also raises questions about privacy and control. Unlike traditional television, where viewing habits remained largely private, streaming platforms operate on continuous surveillance. And unlike social media, where data collection is often explicitly acknowledged, streaming platforms tend to frame these processes as part of user experience design.
The Privacy Problem Few Viewers Think About
Most viewers know that social media platforms such as Facebook, Instagram and Twitter hoard your data for targeted ads and other profitable ventures. But do streaming services, that cater to so many more viewers, also profit from your digital information? While this development may seem incremental in many ways, it represents a new phase in the evolution of how platforms extract value from the activities of their users. User interactions are no longer simply tracked to measure audience size; they are continuously analyzed through AI-driven systems that shape content visibility, influence consumption patterns, and generate economic value. As a result, viewers become both consumers of content and producers of the data that underpins platform profitability.
In Canada, the Personal Information Protection and Electronic Documents Act (PIPEDA) requires organizations to obtain meaningful consent for the collection and use of personal information. However, as recommendation systems become increasingly sophisticated, questions remain about whether users fully understand how their viewing behaviour is analyzed, profiled, and monetized. While privacy laws focus on transparency and consent, AI-driven streaming services generate value through continuous data collection and behavioural prediction, raising concerns about whether existing protections are keeping pace with technological change. These concerns are reflected in public opinion. According to the Office of the Privacy Commissioner of Canada, 91% of Canadians are concerned about their personal information being used to create marketing profiles. At the same time, only 47% of Canadians rate their knowledge of their privacy rights as good or very good, a figure that has steadily declined since 2020. This suggests that many users may be concerned about privacy but lack a clear understanding of how their data is collected, analyzed, and monetized by digital platforms.
What Are We Really Trading for Convenience?
As viewers, we are often told that personalization improves our experience, that smarter recommendations mean less time searching and more time enjoying content. In many ways, this is true. AI-powered recommendation systems can help users discover new content, create more relevant viewing experiences, and make increasingly large streaming libraries easier to navigate. But these benefits come with trade-offs.
Every recommendation algorithm is there to optimise for something else: to keep you within the system, to encourage you to be more active, to be more predictable, to be more valuable within the system. And the system always knows best.
So, Who Benefits the Most?
The rise of AI-driven streaming platforms has blurred the line between serving audiences and studying them. As platforms such as Paramount+ become more dependent on data-driven personalization, viewers are no longer simply consuming content; they are continuously generating data that can be analyzed, monetized, and used to shape future experiences. The result is a system that offers genuine convenience while demanding unprecedented access to personal behaviour.
Whether that trade-off is worthwhile remains an open question. For some viewers, highly personalized recommendations may justify the collection of their data. For others, the growing ability of platforms to monitor, predict, and profit from audience behaviour may represent too high a privacy cost. As AI becomes increasingly embedded within streaming services, the challenge will not be choosing between personalization and privacy, but determining how much privacy we are willing to exchange for convenience.
Infographic by Eirian Thompson-Brown.
University of Toronto Student
Eirian Thompson-Brown is an undergraduate student at the University of Toronto pursuing an Honours Bachelor of Arts in New Media Studies and Media Studies. Her interests include film, digital media, web and game development, and exploring how emerging technologies such as artificial intelligence are shaping society, culture, and communication.
Sources
Adobe Communications Team. (2024). Behind the scenes — How Paramount+ used Adobe Firefly generative AI in a social media campaign for the movie IF. Adobe for Business.
https://business.adobe.com/blog/behind-scenes-how-paramount-used-adobe-firefly-generative-ai-social-media-campaign-movie-if
Baumgartner, J. (2024). Content creators must be “sensible” on how AI is used, Paramount CTO says. Light Reading.
https://www.lightreading.com/ai-machine-learning/content-creators-must-be-sensible-on-how-ai-is-used-paramount-cto-says
Colos, L. (2024). Paramount Global: AI use cases 2024. Pitchgrade.
https://pitchgrade.com/companies/paramount-global-ai-use-cases
Federer, K. (2024). How Paramount and Spaceback are unlocking CTV advertising for all. Spaceback.
https://www.spaceback.com/post/how-paramount-and-spaceback-are-unlocking-ctv-advertising-for-all
Frankel, D. (2025, January 23). Streaming took a bite out of cable’s TV usage share in December. StreamTV Insider. https://www.streamtvinsider.com/video/streaming-took-bite-out-cables-tv-usage-share-december?=
Google Cloud. (2024). Paramount Global built its streaming platform on Google Cloud. Google Cloud Blog.
https://cloud.google.com/blog/products/media-entertainment/paramount-global-built-its-streaming-platform-on-google-cloud
Google Cloud Consulting. (2023). Video content summarization and metadata enrichment for streaming platform. ZenML LLMOps Database.
https://www.zenml.io/llmops-database/video-content-summarization-and-metadata-enrichment-for-streaming-platform
Mulligan, T. (2024). Why the ‘New Paramount’ is placing AI creation at its core. MIDiA Research.
https://www.midiaresearch.com/blog/why-the-new-paramount-is-placing-ai-creation-at-is-core
Office of the Privacy Commissioner of Canada. (n.d.). Consent. Government of Canada. https://www.priv.gc.ca/en/privacy-topics/business-privacy/collecting-personal-information/consent/?utm
Office of the Privacy Commissioner of Canada. (2025). 2024–2025 public opinion research on privacy issues: Final report. Government of Canada https://www.priv.gc.ca/en/opc-actions-and-decisions/research/explore-privacy-research/2025/por_ca_2024-25/?utm_source=
Pajkovic, N. (2022). Algorithms and taste-making: Exposing the Netflix Recommender System’s operational logics. Convergence (London, England), 28(1), 214–235. https://doi.org/10.1177/13548565211014464
Paranandi, S., & Tripathi, A. (2024). Paramount+: A streaming powerhouse with limitless entertainment. Google Cloud Blog.
https://cloud.google.com/blog/products/media-entertainment/paramount-global-built-its-streaming-platform-on-google-cloud
Paramount Global. (2025). Annual report pursuant to Section 13 or 15(d) of the Securities Exchange Act of 1934 (Form 10-K) for the fiscal year ended December 31, 2024. U.S. Securities and Exchange Commission. https://www.sec.gov/Archives/edgar/data/813828/000081382825000005/para-20241231.htm?
Poell, T., Nieborg, D., & van Dijck, J. (2019). Platformisation. Internet Policy Review, 8(4), 1–13. https://doi.org/10.14763/2019.4.1425
Poell, T., Nieborg, D. B., & Duffy, B. E. (2023). Platforms and cultural production. Polity.
Platform Society. (n.d.). Oxford University Press. Platform Society – University of Toronto
Rudyl.ai. (2024). AI at Paramount Global.
https://www.rudyl.ai/ai-research/companies/paramount-global
Torres, R. (2025). How Paramount flipped the script on AI adoption. CIO Dive.
https://www.ciodive.com/news/paramount-AI-adoption-roadshow-lakshman-nathan/807836
Van Dijck, J. (2014). Datafication, dataism and dataveillance: Big Data between scientific paradigm and ideology. Surveillance & Society, 12(2), 197–208. https://doi.org/10.24908/ss.v12i2.4776
Winslow, G. (2024). Paramount advertising brings AI-generated ads to Pluto TV, Paramount+. TVTechnology.
https://www.tvtechnology.com/news/paramount-advertising-launches-ai-powered-paramount-ads-manager
ZenML. (2023). Video content summarization and metadata enrichment for streaming platform. LLMOps Database. https://www.zenml.io/llmops-database/video-content-summarization-and-metadata-enrichment-for-streaming-platform
Zuboff, S. (2020). The age of surveillance capitalism : the fight for a human future at the new frontier of power (First edition.). PublicAffairs.
