Lecture Notes

COMM 2P50 What to Do  Forum posts (35%) o Post: Every Friday by 7 pm o Response: Every Monday by 7 pm o Requirements:  Respond to prompt and apply course content  At least 1 citation from readings or lecture of that week  500-850 words o Response:  Respond to one post  Full sentences, demonstrates reflection  200-550 words  Follow APA Style  Citations: o Citing lecture in reference list:  Last name, F. M. (Year, Month Date). Title of presentation [Lecture notes, PowerPoint slides, etc]. Publisher. URL  EXAMPLE: Smith, C. (2022, October 13). AI and machine learning demystified [PowerPoint slides]. Sakai. http://www.someurl.com o Citing Lecture in-text:  Paraphrasing: Professor's last name, Year. (Chen, 2021)  Direct quote: Add slide number. (Chen, 2021, slide 6)  MIDTERM (20%) o Open date: November 29 th o Close date: Dec 5 @ 11:59pm o What to do:  Provide answers weighing different perspectives from the lecture, readings, and supporting findings/evidence from reputable sources. When citing findings/evidence, include a reference list, unless you are drawing from experiences/observations that are not publicly available.  You are expected to include academic sources, both external and internal in all your answers.  All sources must be paraphrased (no direct quotes allowed).  Your reference list does not count towards your page limit.  Multimedia Proposal (15%) o Due date: Feb 17 th , 11:59pm o What to include:  Short description of your project  What is your project about?  What issues/topics/themes will you be discussing?

 Briefly explain the importance of your project – i.e., answer the “So what? Why should we care” question? (Provide evidence to substantial your claims or arguments)  Short description of your choice of digital media platform  State your choice of digital media platform  Explain your choice of platform and why is it appropriate for your project  • Short description of TWO course concepts.  What TWO course concepts will you be drawing from? Define the concepts in your own words.  Explain how each course concept is appropriate for your project  Find ONE example to demonstrate/support/challenge/build your argument for EACH of the TWO course concepts (aka you need one example for each concept).  Short description of how you intend to execute your project (this will help you think through how much time, effort, and resources you will need for your project)  What do you need, where would you look for your evidence/examples, what equipment/software do you need, do you need tech support?  Include a brief timeline or steps of project execution  If applicable, provide a brief description of any concerns, questions, or challenges. o Length:  Proposal of around 6-8 pages.  Your reference list or cover page do not count towards your page limit.  Format: Double-spaced, 12-point, Times New Roman. APA citation style. First or third person writing style.  Multimedia Project (30%) o Due: April 3, 11:59pm o What to do:  Identify a course-related topic, issue, debate, controversy, or problem and build a multimedia project around that issue—helping your peers understand the implications of that topic, issue, debate, controversy, or problem for how we think about digital media in our society.  Your multimedia project should be built around a course theme or issue (e.g., algorithms, misinformation, influencer marketing, etc) and go beyond course materials to bring in evidence, examples, and perspectives from outside sources to add insight and depth to your topic of choice.  Must receive link to project (YouTube videos, Podcasts, Social Media posts, Website, E-Magazine, PPT with voice-over) o Length:

Module 1 – Week 2  New Media definitions and reasons for being o 1 st definition (scholars): Emerging communication and information technologies undergoing a historical process of contestation, negotiation, and institutionalization. o 2 nd definition (more fitting?): Media we do not yet know how to talk about.  Digital media technology is constantly evolving and developing – its purpose, role, and effects are not yet fully understood o Reason for being: Because we don’t yet fully understand the role of digital media technologies on society – the impact of their affordances & design choices as well as the impact of those who use these technologies to their benefit  Triple revolution: Nothing has matched speed of growth around the world o Triple Revolution that resulted in a corpus amount of information being shared, repurposed, and re-shared in digital formats  Broadband internet deployment  Mobile connectivity  Social media sites & Messenger apps o Why was internet adoption so rapid and widespread – how did it get to hyperconnectivity  Hands-off policies preserved the competitive free market of the internet  Transferred control to private sectors  Telecommunications act of 1996 (USA): Internet = information service o Helped preserve the free market nature of the internet o Incentivized to continue working on internet o Modern version of ‘gold rush’: Everyone rushed to develop it o Rapid production made computers and tech cheaper/accessible to everyone  Improvements in processing speed & power  Increased bandwidth & RISE of broadband (shifted our expectations of how FAST content should be delivered) – Dial-up connection; took 2 hours and could not use internet and phone at the same time  Could stream movies now, improved internet consumption  Downloading and listening online  Watch on demand  Shift towards wireless – Wi-Fi, laptops and mobile phones  differentnt ways to learn  Interconnected network of different service providers o They could all communicate with each other (not fragmented)  Network of networks allowed information to be connected

 Desire to stay connected  Less connected by wires, more mobile and accessible  Server farms and cloud computing leading to email, social media, and online videos becoming viable  Cloud saving – everything could be accessed online  Supported apps  “Killer” apps building on top of the internet infrastructure  Email – most important app for internet users for widespread adoption and use  Where are we now? o In 1 minute in 2019:  Global online users sent over 41.6 million mobile messages  2.1 million snapchat messages  Pandemic and stay -at -home mandate has increased all digital media use  Canada’s New Media Literacy Landscape: 2021 stats o There were 35.63 million internet users in Canada as of Jan 2021 o The number of mobile connections in Canada in January 2021 was equivalent to 98.0% of the total population. o There were 32.20 million social media users in Canada o The number of social media users in Canada was equivalent to 84.9% of the total population o Internet users aged 16-64 (in Canada) spend almost 2 hours on social media every day: YouTube, Facebook, and Facebook messenger used most  Internet Skills and why they matter o Hard skills: Set of technological skills related to using and navigating the internet and information o Soft skills: Critically engaging with digital tools, leveraging its features, and working within their confines.  It is also about thinking critically about conceptual approaches to the uses and implications of digital media on society. o New media literacy: Knowing how to access, analyze and evaluate online information and knowing how to create, use and communicate content/ideas using digital media technologies. o We get our information online all the time and spend many hours on digital media but it is not always trustworthy, could spread misinformation, manipulate results, present a bias, become an echo chamber from the algorithms, affect mental health o With photoshop and other tools, it is hard to tell what is real/fake Assigned Reading: 6 Internet Skills and Why They Matter Citation: Hargittai, Eszter, and Marina Micheli, 'Internet Skills and Why They Matter', in Mark Graham, and William H. Dutton (eds), Society and the Internet: How Networks of Information and Communication are Changing Our Lives, 2nd edn (Oxford, 2019; online edn, Oxford

 Critical Thinkers able to situate information within broader social, cultural, and political contexts  Effective creators & communicators to create, share, and circulate content for political and civic purposes  Agents of political and social change to make significant contributions to civic life. Module 1 - Week 3: Social Shaping of Technology  Concepts related to Digital Media o Current perspectives & why we need a multi-perspective approach  Technological Determinism  Social determinism  Networked individualism  What to do with all assignments: o Make sure you define terms o Provide an example that helps you explain, elaborate, support, or understand the concept o Make connections between and among concepts – how do they compare with each other o Multi-perspective approach to understanding Digital Media  Important issues tied to digital media cannot be addressed from any single conceptual perspective alone.  Classic Positions About New Media o It is a blessing:  Give power back to the people  Bring democracy  Develop apps to solve problems and bridge educational/public health gaps o Curse  Cause of all social problems (misinformation, mental health)  Leads to moral panic – needs to be regulated o Not worthy of notice (uncommon)  Apps are tools and a fad  Will be replaced by new tech so ‘trends’ of today are not a concern  Technological Determinism o There is only one best way of doing something  Tech is on a predetermined path towards the one best way o It will have a predictable set of outcomes (positive or negative)  Direct cause and effect that media has on society o Tech is the main cause of social structure/cultural values o Uncritical and unquestioning faith in tech to do either good or bad  Techno-optimism or techno-pessimism o Ex. Tinder is promoting a promiscuous lifestyle among young people

o Ex. Tech could dismantle hierarchal methods of communication – could lead to democracy  Social Determinism o Humans made specific tech to reach goals or social purposes  Economical, political o Prioritizes society ABOVE technology  Technology could be used by society to effect outcomes, but media is not the cause (people are) o Creation new media was built because of society (to stay in touch). Any other impacts are due to our social interactions o Counter: People create a perfect life on social media that doesn’t happen without it  Social Construction of Technology o Both Technological and Social Determinism are flawed o There is interaction between structure (tech) and culture/society o Digital media is used in social contexts, but using the media shapes the social context they are in  Not deterministic o Understanding the technical, economic, political, gendered, geographical, and other social factors shaping technologies can foster more equitable outcomes o Society affects media creation, and media affect progressing society o Understand the structure and culture of new media (The Internet)  Trajectory and growth are not by chance  Combination of political, technological, and social impact that make current digital media technologies what they are today o Structure of New Media  New media vs. traditional media (TV, radio…)  Affordances/defining features: set of possible actions a person can perform on an object, which affects how people use them.  They have social implications (design of email = fast speed of communication = reply ASAP)  Twitter hashtag: Organize by key-words, things are easy to search for – helps raise awareness of shared grievances, organize movements more easily, helps like-minded people find each other  React button: Provide peer feedback to maintain relationships (pressures people to cultivate likes)  Networked Individualism o The structure/design forces that set digital media apart from traditional media o Design choices in the hardware, software, and connectivity have fostered networked individualism  PC: personal, user-friendly, customizable, private, and asynchronous  Encourages content creation and sharing instead of passive consumption

 Infuses online culture with the spirit of individualism through looser networks o Participants  Active users who post materials meant to influence and help others (e.g. blogs, support groups)  the so-called Fifth Estate in civic life  Moral Panic o Issues are often LESS about the technology and MORE about the changing moral and value boundaries of our society o Contentious topic: Sex  Blurs boundary between private and public  Explicit content easily accessible, especially to those who were once forbidden from viewing them  Example: Snapchat & sexting  Result: Call for more regulation & restriction o Contentious topic: Cyber-bullying and toxic behaviour  Being anonymous removes social restrictions  Underlying assumption is that humans are inherently bad and will behave poorly when no one is looking  De-individuation (decreased sense of personal responsibility that inspires disinhibited behavior and impulsiveness) versus group norms & social cues  Group/community culture matters  Example: Anonymous texting  Result: Censorship, regulation, surveillance Module 2, Week 5: Trapped by our own digital footprints  Filter Bubble and Echo Chamber o Filter bubble: Created when news and information are filtered by algorithms, keeping people within a bubble of similar views o Echo-chamber: When we surround ourselves with information and viewpoints that supports our worldview (note: we can have echo-chambers without algorithms) o Dangers of algorithmic filters (filter bubble) + social filter (echo-chamber): Not exposed to diversity of viewpoints, leading people to believe in different worldviews based on our information sources o Example: 2016 US Elections laid bare the implications of living inside an echo- chamber (SNL sketch calls it The Bubble)  Algorithmic Filter Bubble o • History of our online behavior - clicks, likes, comments, shares, search terms – is what powers the algorithmic engine

o An algorithm: An effective method expressed as a finite list of well-defined instructions for calculating a function o Starting from an “initial input” (start), the instructions describe a computation that when executed will proceed through a finite number of well-defined successive states, eventually producing an output and terminating at a final ending state. It transforms inputs into outputs o Plain English: A process or set of rules a machine (e.g. computer) follows to achieve a particular goal. o Algorithms make choices on behalf of users by predicting what users want to see & are created to maximize profits (e.g. Ad revenue) o An array of options are recommended, and a choice is made by the user that is then fed as new knowledge to train the algorithm into predicting better and more relevant content to the user o This leads to a chicken-egg question: Did you click on something because you were inherently interested in it, or did you click on it because it was recommended? o This creates a feedback loop, where the output of the algorithm becomes part of its input o This results in similar recommendations reinforcing the popularity of some types of content over others o Algorithmic amplification: when some content becomes popular at the expense of other viewpoints o Users encounter news stories and information that are tailored to their interests BUT decrease a person’s exposure to diverse opinions and perspectives (an unintended consequence?)  Is your news filtered? o Self-selected personalization: People active users of media. They have always made decisions to personalize their news use o Limited control: Users are largely unaware of how news personalization algorithm works or how their data is being tracked for news personalization o People’s access to info are shaped by processes which are largely hidden to them o Implication: How do we un-filter our news if we don’t know what is being filtered?  Evidence of Filter Bubble and Echo Chamber: o Examined web-browsing histories for 50,000 Internet users who regularly read online news o Echo-Chambers  Individuals generally read content from news outlets that are ideologically like each other  Users who regularly read partisan articles are almost exclusively exposed to only one side of the political spectrum

 Social Shaping of Technology o If algorithmic filtering is only part of the problem, then what makes up the other parts?  Politics becoming more “affectively polarized” in Canada?  Personal biases? (e.g. homophily, confirmation bias?)  Information as identity affirming rather than truth-seeking? o Social shaping of technology: Dialectic model between technology & social users.  Actions implemented through the algorithm, but through the interactions that take place between the algorithm and digital media users.  Online behavior is shaped by algorithms, but also by the beliefs users have about what the technology is best suited for, how they use it, and the norms they develop when using the technology o From Searching through filter bubbles, echo chambers & disinformation (Dutton, Reisdorf, Blank, Dubois, & Fernandez, 2019)  Cross-national comparative study on Internet & Social Media use conducted in 2017  Challenges filter bubble perspective: Search plays a major role in shaping access to information – but it is not deterministic  Only one of the multiple media people use to find information  Users are exposed to a diversity of viewpoints through a diversity of media found via search  Challenges echo chambers perspective: Internet users are exposed to diverse viewpoints both online and offline  Users are exposed to contrasting views online & not everyone silences others who disagree with them  Use multiple platforms and information sources = design of any single platform becomes less meaningful  Does not negate the role of confirmation bias nor does it mean internet users trust all sources equally BUT it does counter claims that digital media is to be blamed for putting people into filter bubbles  Formation of public opinion about political issues and current affairs is an on-going social process that is not limited to a single digital media platform or social interaction  Keep in mind: Use of search engines & social media to learn about politics and current affairs is NOT central to our use of digital media  How often do you use social media or Google search to look for information about political events and current affairs?  We use digital media mostly to socialize and to be entertained – algorithms and search personalization will likely show us content unrelated to and beyond our political beliefs. o Multiple Perspectives

 Technological-deterministic perspective: Systems that personalize information through algorithms will predictably trap users in a filter bubble - e.g., filter bubbles  Social-deterministic perspective: Individual’s pre-existing political attitudes and beliefs will predictably lead users to seek out information that confirms their biases – e.g., echo chambers  Social-technical shaping perspective: Outcome of information exposure is far less predictable than we might expect based on either the technology or individual preexisting beliefs. Information exposure is shaped by technical, social, economic, and cultural factors. Focus should be on what users’ actions under different contexts of use (aka not some pre- determined outcome). Module 2, Week 6: Politics of Big Data  Is code just code? o We are now turning to algorithms to identify what we need to know is as momentous as having relied on credentialed experts, the scientific method, common sense, or the word of God o Development of a new “knowledge logic”: Set of presumptions about what knowledge is and how it is organized with implications for what we think & believe about the world o What is known and what can be known – possibilities that are not known to us because algorithm determines what is shown to us o Remember last week’s lecture?  Most us have little knowledge or are even aware of the calculations involved in determining content results  Most users are not even conscious of the impact of algorithm (it’s the propriety magic that gives tech companies power over our data. And it works seamlessly behind the scenes)  Public Relevance Algorithms o Let’s break down Gillespie’s (2014) public relevance algorithms  They “select what is most relevant from a corpus of data composed of traces of our activities, preferences, and expressions.”  This has important and serious implication on how we find information, on “what there is to know” and “how to know it” o Algorithms as communication technology, much like broadcasting and publishing technologies. o Six dimensions:  Patterns of inclusion  Understanding what data gets in what data gets left out  Cycles of anticipation  Knowing users in order to make predictions about their behavior  The evaluation of relevance

o Commercial aims  When algorithms come to represent legitimate knowledge, new commercial endeavors are made to fit them (think of SEO)  Facebook News Feed: mixing the social (e.g., status updates, friends’ activities) and commercial (e.g. banner ads, third-party sites) in a single space o • Epistemological premises  Ideas about “quality” information are embedded in algorithms  “evaluations preformed by algorithms always depend on inscribed assumptions about what matters, and how what matters can be identified” o Algorithms are not stable (e.g. Google & A/B testing)  The promise of algorithmic objectivity o More than mere tools, algorithms are also stabilizers of trust, practical and symbolic assurances that their evaluations are fair and accurate, and free from subjectivity, error, or attempted influence. But, though algorithms may appear to be automatic and untarnished by the interventions of their providers, this is a carefully crafted fiction. In fact, no information service can be completely hands- off in its delivery of information… Yet it's very important for the providers of these algorithms that they seem hands-off. o Algorithm bias  Biases are encoded within algorithms  Amazon scraps recruiting tools that showed bias against women candidates  Story of Dr. Latanya Sweeney who has done research on online ad discrimination  Image cropping algorithms  Entanglement with practice o Why entanglement with users rather than effect of algorithms on people?  Users’ decisions affect algorithms  What people do or not do is already a kind of calculation o Algorithmically recognizable  Redesigning expression so as to be recognized by search algorithms  Seeking to be noticed: e.g., hashtags  Seeking to evade notice: e.g., misspelling of words to evade censorship o Domestication  Technologies embedded into everyday routines, in the home, etc.  Users ‘make algorithms their own’ (for relevance and convenience e.g. Amazon; Ads)  Users have agency too o Knowledge logics  Algorithms offering relevant knowledge also offer ways of knowing  “What works” can seem to equate with “what’s right”

 Production of calculated problems o “Calculated publics” = algorithmically generated publics that are assembled by algorithms, representing a certain kind of (constructed) public shaped by unspecified criteria (e.g. recommendation systems) and are not user-led  Amazon recommendation & Facebook Groups  News websites calculated publics  Trending topics/communities/hashtags you may like suggested to you by a website  Common Perspectives o “Code is just code”/ “it is a step-by-step instruction”  “Algorithms might be unbiased in its execution but that does not mean bias has not been encoded within it” (e.g. code itself isn’t bias. It does what it is supposed to do)  Algorithmic identity (therefore implications of that code) ◦ Gives companies an excuse to minimize responsibility – the need for ethical algorithms o “Do we have a choice?” o Algorithms are everywhere. There is nothing we can do about it”  We can be cognizant of its limitations & its impact on knowledge (what is known to us and what can be known to us)  Recognize that we have a choice to be active in searching for different perspectives; to dig deeper on Google Search; to question trends/recommendation lists and FB News Feed o Accurate algorithmic prediction means more relevant content/useful ads  Free will vs. prediction? (e.g. data doppelgangers – versions of our past, present, and future selves)  Does it take the work of ‘thinking’ & ‘judging’ away from humans?  Will/has it change(d) our focus to ‘sense of general direction’ & outcomes instead of understanding cause to problems?  Data divide between us and our own data  Always an issue of blind spot and issues of representations because not everyone can participate in the datasets the same way; so algorithmic predictions are never truly objective or fair (esp when it comes to “solving” social problems) Module 2, Week 7: Implications of Algorithms  How algorithms reflect our biases: o Algorithms play a critical role in our lives because digital media is such an integral part of our lives  Affects the answers we get, the knowledge we can access, the content we see, our professional networks, movies we should watch, types of advertising ads we get, etc.

 Algorithms designed to maximize profits might exclude certain groups of people unfairly  Biases in the code because programmers/engineers may have inherent biases o AI systems learn to make decisions based on training data  “Bias in, bias out”: Dataset to train machine learning algorithms is already biased (Amazon hiring algorithm)  Flawed data sampling overrepresenting or underrepresenting groups  “What gets chosen is usually whatever is easiest to quantify rather than the fairest” o Limits of algorithms: Silly to assume that algorithms can be “fair” (Heilweil, 2020)  Algorithms are not great at nuances or subjectivities  Social issues don’t necessarily have an objective solution  Recognize that some algorithms come with a high risk of abuse (e.g., predicting future crime based on online behavior etc.) o “Allocative harms” versus “representational harms” (Bornstein, 2017)  Need to pay more attention to how minorities are presented in search results  Why is this statement important? --- representation is “often overlooked in discussion of fairness” because “if you can't study something in formal [mathematical] terms, its existence is not as legitimate as something you can turn into an equation or an algorithm”  Goes back to the argument that what gets included into an algorithm is often what can be easily quantified rather than nuances o Two outcomes  Negative: Reinforcing status quo and further deepening inequalities  Positive: Expose and force society to confront historical biases and inequalities & force us to reassess our definitions of fairness and equity  Positive: Disregard variables that do not accurately predict outcomes compared to humans  Fixing biases in machine learning algorithms o Is it realistic to expect a completely unbiased AI algorithm?  Training data reflects biases and inequalities in our society  Algorithms checking dataset to identify and remove biases are also designed by humans  But that doesn’t mean we shouldn’t try to minimize biases o What are the best practices of AI bias minimization?  Be aware of contexts in which AI can help correct for bias and those in which there is high risk for AI to exacerbate bias  Establish processes and practices to test for and mitigate bias in AI systems

 Engage in fact-based conversations about potential biases in human decisions (Positive impact of AI)  Fully explore how humans and machines could best work together  Invest more in bias research, make more data available for research while respecting privacy, and adopt a multidisciplinary approach  Invest more in diversifying the AI field  Data capitalism and commodification o Business decisions determine how algorithms are used in advertising  Profile-based algorithms: Attempt to uniquely identify users by linking their online activity to their offline life e.g., DoubleClick  Interest-based algorithms: Attempt to mine the keyword search history to infer purchase intentions e.g., Google o Profile-based algorithms require immense databases of uniquely identifying information compared to interest-based algorithms o Google acquired DoubleClick and merged both advertising approaches o Data capitalism: Commodification of data enables a redistribution of power  A shift from sale of goods online (e-commerce) to the sale of audiences (users are the product to be sold to companies and advertisers) o Data commodification: Data is treated as commodity to be sold and circulated  The rise of data brokers - companies that collect data from multiple sources & platforms, and then repackage them to be sold to advertisers  Can have negative consequences that impacts pricing, lending decisions, and interest rates o Not just data-broker but tech giants who own and manage user data  Own their proprietary ad networks, tracking ads, and technologies that mine user data that they can then monetize e.g., Google & Facebook  You (the consumer) are the product o Ad tracking tools enable advertisers to collect data about us; helping them determine engagement metrics and purchases behavior. o Websites earn ad revenue from displaying the ads to you o Tracking URLs: Typical in Pay-per-click campaigns, email or ad campaigns o Tracking pixels  Typical when companies want to know how users are interacting with the ad from the initial click through the final purchase.  Companies use this to track users through every step of their sales journey  Ads on Facebook typically use this method o Cookies  Driving force behind most retargeting/remarketing ads  Typical when companies want to create a personalized experience for users based on their previous interactions with the company e.g., abandoned cart email or abandoned cart ad