HMCS Vancouver Pre-Deployment Brief - May 2024 (Web Version).pptx
Final Report.pdf
1. FINAL REPORT
NICHOLAS THOMSEN
THOMAS EDISON STATE UNIVERSITY
SOM-7020-MB900: SOCIAL MEDIA MARKETING
PROFESSOR DR. DAVID J. CASTLE, PH.D.
AUGUST 20, 2023
2. TOPIC SELECTION
ARTIFICIAL INTELLIGENCE
• Artificial Intelligence (AI) is a method of making a computer or software think
intelligently like the human mind. It studies the patterns of the human brain and
analyzes the cognitive process such as perceiving, reasoning, learning, interacting
with an environment, problem-solving, and even exercising creativity. (McKinsey &
Company, 2023) AI systems work by merging large data with intelligent, iterative
processing algorithms. This combination allows AI to learn from patterns and
features in the analyzed data. Each time an Artificial Intelligence system performs a
round of data processing, it tests and measures its performance and uses the results
to develop additional expertise. (Duggal, 2023) AI is implemented through machine
learning and deep learning.
Image: Adobe stock
3. ARTIFICIAL INTELLIGENCE IN
MARKETING
• AI serves many functions in marketing. Social listening improvements result from
user-entered queries, suggesting a vast range of terms to be tracked to help quickly
discover audience insight. Monitoring keywords and consumer sentiment makes it
easier to determine what the target audience is interested in. Automation improves
efficiency and ensures the delivery of the right message at the right time to the right
person. Vast amounts of data make formulating a target market easier, and support
features improve customer interaction and support. 24/7 data collection and
monitoring provides real-time data to users for reputation management in the event
something goes wrong. Language processing provides a multilingual advantage in
providing customer care. (Chacko, 2023)
4. DEFINITIONS/KEY THOUGHTS
AI can be broken down into weak and strong AI.
• Weak AI: Performs a specific task and is limited to that task only. An example of
weak AI is Amazon’s Alexia and Apple’s Siri voice assistants.
• Strong AI: Systems that possess human-level intelligence or higher across a wide
range of tasks. This AI would be capable of understanding, reasoning, learning, and
applying knowledge to solve complex problems in a manner similar to human
cognition. (Duggal, 2023)
AI can further be broken down into various types (Duggal, 2023)
• Purely reactive: Have no memory to work with and specialize in just one field of
work. IBM’s Deep Blue that was able to beat the world’s best chess player in 19
moves.
• Limited memory: Machines collect previous data and continue adding it to their
memory. DeepMind’s AlphaZero was able to teach itself chess in four hours and win
or draw each of 100 games against the world champion chess program Stockfish 8.
(Kasparov, 2017)
• Theory of mind: The AI can understand thoughts and emotions, as well as interact
socially.
• Self-aware: Will be intelligent, sentient, and conscious.
These last two do not currently exist.
5. IMPACT
• Since its inception, AI has made great strides in its capabilities and functions.
Starting off performing simple tasks, AI has developed into a human aid product
spanning extensive fields. AI is now prominent in social media and marketing
providing a great system for data collection and analysis and even the generation of
images, video, and language. With further development, there is no limit to its
potential uses. Marketing has seen improvements in ROI, interactions, audience
targeting, campaign planning, media generation, and customer support functions, to
name a few.
Image: Future of Life Institute
6. ANALYSIS
HISTORY
• The idea of Artificial Intelligence isn’t new. The idea can be traced back as far as
1305 when a polymathic by the name of Ramon Llull published “Ars generalis
ultima,” including the concept of producing logical deductions by mechanism rather
than human thought.
• The term artificial intelligence was coined in 1955 in the publication of “A Proposal
for the Dartmouth Summer Research Project on Artificial Intelligence “(Rosso,
2017) Just now, almost 70 years later, AI is starting to show up in more business
markets performing a variety of functions.
• From 1957 to 1974, AI flourished due to improvements in computer data storage,
speed, cost, and accessibility. Machine learning algorithms improved, as did
knowledge of how to properly apply them.
Image: Engineers Garage
7. ANALYSIS
HISTORY
• Though improvements were made, dwindling patience and funding slowed progress
significantly for ten years. Computer technology couldn’t store and process data fast
enough to exhibit human-like intelligence.
• After a boost in funding and expanding algorithms, AI once again gained popularity
and led to the development of Deep Blue, IBM’s chess-playing AI that was able to
beat the reigning chess world champion Gary Kasparov in 1997.
• Development of various AIs continues in vision and language, banking, marketing,
and entertainment. (Anyoha, 2017) It wasn’t until recently, with our increased
computing power, global internet connection, and cloud-based computing that we
have been able to realize the potential of AI.
Image: iStock / Getty Images Plus
8. ANALYSIS
PROS
• To better interact with customers, AI can fine-tune content, learn customer preferences,
quickly pick up on changes in customer behavior, and can offer relevant
recommendations.
• To better support marketers, AI can predict customer spending patterns, perform
market research, and provide real-time customer support when a live agent isn’t
available.
• Data collection and recognition can leverage billions of customer interactions to
develop messages that will motivate customers to engage and act.
• Automated processes, such as data collection, free up more time for marketers to
create content, strengthen brand messaging, and develop campaigns. This improves
overall efficiency by removing many time-consuming administrative and manual
tasks. (Starita, 2023)
9. ANALYSIS
CONS
• Reliable prediction and analysis can be uncertain. Reliable sentiment analysis is hard
for AI, as are visually recognizing objects and folding clothes for AI-powered robots.
• Creating an AI to completely replace a human is extremely difficult. Instead, most
are used to support the human user.
• AI can only work within the constraints of the system. AI functionality is limited to
the inputs it receives, and the algorithms it uses. Lost data from one platform will
significantly impact the system as a whole.
• AI isn’t human. It is currently unable to have emotion, feelings, and subjective
thinking, limiting its ability to form emotional connections as a human would.
(Roetzer, 2021)
• Unethical uses have a significant impact. Some aspects of AI have made it difficult
to determine if something is fact or fiction, making it easier than ever to deceive
people.
10. ANALYSIS
EXAMPLES OF USE
• The basis of this system started in 1992 with the concept of “collaborative filtering,”
using rules to filter emails that could then relate that opinion and behavior to others.
Further development included data from user rating systems and machine learning to
aid in predictions. (Jannach et al., 2021)
• Recommendation Systems are software tools in applications or websites that suggest
information that might be of interest to the end-user, taking into account various
types of knowledge and data, such as the user’s preferences, actions, tasks, and
contextual information. In most cases, these systems use computational methods to
analyze users’ past actions and decisions, along with other user-related or task-
related information, to offer useful and usually personalized recommendations.
Image: Medium.com
11. • Three basic types of recommendation systems now exist.
• Collaborative filtering is an approach where recommendations were given by others who
have similar preferences in the past, but who already experienced an item or product yet
unknown to the current user. Collaborative filtering systems require users to express opinions
on products or items. They collect the opinions of users and recommend items based on
people’s opinions similarities.
• Content-based filtering works with data that the user provides, either explicitly (rating) or
implicitly (clicking on a link). Based on that data, a user profile is created, which is then used
to make recommendations to the user. As the user provides more inputs or ratings or takes
action on the recommendations, the engine becomes more and more effective and accurate.
Content-based recommendation systems try to recommend products or items similar to those
a given user has liked in the past, whereas systems designed according to the collaborative
recommendation model identify users whose preferences are similar to those of the given
user and recommend products or items they have liked. Content-based filtering techniques
recommend items based on a comparison between the content of the items and a user profile.
• Hybrid filtering combines collaborative and content-based filtering in several different ways.
This is done by performing predictions for each and then combining them, by adding
content-based capabilities to a collaborative filtering technique or by unifying both
techniques into one model. The main goal behind hybrid filtering is that a combination of
algorithms will provide more accurate and effective recommendations than a single
algorithm as the disadvantages of one algorithm can be overcome by another algorithm. This
is found in various forms across the internet such as Netflix video recommendations and
Amazon product recommendations. (Varshini, 2020)
12. ANALYSIS
EXAMPLES OF USE
• The investment management company Vanguard, which manages $7.7 million in
global assets, was looking to find a better way to ensure their customers would get
the content they want and need. Primarily focusing on B2B sales with retirement
plan sponsors, Vanguard uses only LinkedIn for social media advertising in the
heavily regulated sector. To gain market share, Vanguard needed to create messages
that would resonate and encourage prospects to learn more about how Vanguard
could help them achieve their goals.
• Vanguard used an AI called Persado to fine-tune LinkedIn messages. Persado uses an
experimental engine to generate combinations of concepts and phrases to appeal to
different customer needs. It has a vast language knowledge base mapped to human
emotion and trained for enterprise communications. It also provides insight into
audiences, which drives data back to the business to understand why customers
engage with specific language elements over others and to build more effective
content.
• Comparing control messages to those generated by Persado showed an increase in
click-through rate of 15.76%. Additionally, the data received provides insight for
future campaigns. (Persado, n.d.)
13. SUBTRENDS
• Continued growth of generative AI. Programs are available to generate products with
a simple request. Chat GPT will craft well-thought-out responses to questions and
tasks. Synthesia will create a video of a person speaking test that the use enters.
Lexica will generate an image with a simple user-entered description. Though
many are fine-tuning their current product, AI startups are expanding into various
markets with new use cases.
• Expansion of embedded AI and UX-focused AI. Primarily used in third-party
systems, this AI makes it easier for employees to search for documents,
conversations, and other resources. This includes language processing such as
Google Translate.
• Stronger compliance and ethics expectations. People are growing more concerned
with what data is being collected, how it’s used, and if it is appropriately secured.
There is a push for AI companies to make their data collection more transparent.
• Continued AI democratization and widespread AI access. Due to the limited
amount of people with the appropriate skills, businesses struggle to apply AI data
efficiently. User-friendly AI is being developed to help under-staffed or under-skilled
businesses utilize the collected data.
• Improvements in computer vision. This Ai makes it easier for computers to better
understand image-based data. Though frequently used in manufacturing, this
technology can be used to answer captcha prompts, something typically used to
ensure a human is present in the interaction. (Hiter, 2023)
14. INTERVIEW
• The expert interview was conducted with Mrs. Elena Bond, the Marketing Science
Lead for Snap Inc. North America. Mrs. Bond attended New York University and
majored in psycholinguistics or the relationship between language and psychological
processes. Realizing she would need an advanced degree to further her career, she
took additional classes in statistics and marketing research leading her to analytics
and marketing. She currently works in the monetization branch of the organization,
focusing marketing efforts on the two main revenue streams for the company,
premium memberships and paid media from advertisers.
• My additional questions to her were who is her target market, how to best reach
them, what metrics are used for determining the success of a campaign, if Snap uses
AI for marketing, and if so, what function it serves, and the benefits and risks to
using AI in marketing. She emphasized that the target market depends on the
product. She gave the Barbie movie as an example, with the core target demographic
being women ages 13 to 34 and a secondary demographic of women ages 35-54 and
men in the same age ranges. After determining the marketing budget, allocation is
determined for various channels, including co-marketing that would best reach those
target demographics on each platform. Included in that decision should be past
marketing metrics performance for each channel as well.
• Snap uses a form of AI to connect an advertisement with a user at the right time with
the right message. Not specifically used by Snap but AI that is common in other
businesses is using AI for target selection by using certain community indicators. AI
provides convenience and efficiency to many aspects of business but is underutilized
in the potential functions it could provide. Due to being human-designed and
absorbing human information, AI is inherently biased and can be easily exploited for
unethical purposes. To maximize benefits and minimize risk, this technological
advancement will require people to learn new ways to utilize AI ethically.
15. THE FUTURE
• The future of AI is uncertain, but it will progress down one of three paths.
• AI may continue to develop unchallenged as it has previously, continuing the
development of current uses and finding more use cases in many more fields.
• AI may be further developed but under strict regulation.
• A stop to AI technology may be put into effect to protect people from it.
• Leaving AI development unchallenged will ultimately generate many more uses for
it. As with any other technology, as it grows and progresses, people will find more
uses for it to make their jobs and lives easier. Unfortunately, if left unchecked, this
will also provide more effective ways to deceive people and violate their trust.
Examples of this are already available. All someone has to do is use Google to
search for President Biden's deep fakes and multiple videos with AI-generated audio
will be found. This is of concern because someone on a computer with the
appropriate skill level can create an audio file of the President of the United States
saying whatever they want.
Image: Getty Images
16. THE FUTURE
• Having noticed these threats and others, the most likely path forward is with some
regulation. Already in existence is a Blueprint for an AI Bill of Rights that aims to
protect people from various harmful side effects of AI. This includes reducing
discriminating algorithms, keeping data private, notice and explanation of system
uses, and allowing individuals to opt out of AI use. The key to this plan working to is
allowing the development of AI without violating individual rights.
(Whitehouse.gov, n.d.)
• A complete stop to AI is possible but does not seem likely. There has always been
fear that AI could develop so far that it would pose a significant threat to humanity.
Think Skynet and Terminator. 30,000 people have recently signed a petition to
immediately pause the development of AI. These aren’t all people that are against
technology but include Apple co-founder Steve Wozniak, Tesla, Twitter and SpaceX
CEO Elon Musk and Turing prize winner Yoshua Bengio. Though just a pause, if
enough evidence exists from knowledgeable people in the field that AI will
essentially take over the world, this could be enough to force a halt of AI generation
for good. (Marr, 2023)
Image: TheConversation.com
17. REFERENCES
• Anyoha, R. (2017, August 28). The History of Artificial Intelligence. Retrieved August 20, 2023, from
https://sitn.hms.harvard.edu/flash/2017/history-artificial-intelligence/
• Chacko, A. (2023, May 9). The role of artificial intelligence in marketing. Sprout Social. Retrieved August 20, 2023, from
https://sproutsocial.com/insights/ai-marketing/
• Duggal, N. (2023, July 27). What is Artificial Intelligence: Types, History, and Future. Simplilearn.com. Retrieved August 18, 2023,
from https://www.simplilearn.com/tutorials/artificial-intelligence-tutorial/what-is-artificial-intelligence
• Hiter, S. (2023, July 12). Top 6 AI Trends 2023. Eweek. Retrieved August 20, 2023, from https://www.eweek.com/artificial-
intelligence/ai-trends/
• Jannach, D., Pu, P., Ricci, F., & Zanker, M. (2021, November 20). Recommender systems: Past, present, future. AI Magazine, 42(3),
3-6. https://doi.org/10.1609/aimag.v42i3.18139
• Kasparov, G. (2017, December 8). AlphaZero AI beats champion chess program after teaching itself in four hours. Kasparov.com.
Retrieved August 18, 2023, from https://www.kasparov.com/blog-post/alphazero-ai-beats-champion-chess-program-after-
teaching-itself-in-four-hours/
• Marr, B. (2023, May 3). Should We Stop Developing AI For The Good Of Humanity? Forbes. Retrieved August 20, 2023, from
https://www.forbes.com/sites/bernardmarr/2023/05/03/should-we-stop-developing-ai-for-the-good-of-
humanity/?sh=61618fd32943
• McKinsey & Company (2023, April 24). What is AI? Retrieved August 18, 2023, from https://www.mckinsey.com/featured-
insights/mckinsey-explainers/what-is-ai
• Persado (n.d.). Vanguard Boosts Conversion Rates 15% by Trusting AI to Strengthen Client Messaging. Retrieved August 20, 2023,
from https://persado.drift.click/vanguard-case-study
• Roetzer, P. (2021, November 8). 6 Disadvantages of AI in Marketing According to Experts. Marketing Artificial Intelligence
Institute. Retrieved August 20, 2023, from https://www.marketingaiinstitute.com/blog/limitations-of-marketing-artificial-
intelligence
• Rosso, C. (2017, February 21). Why AI is Trending Now. Medium.com. Retrieved August 18, 2023, from
https://medium.com/@camirosso/why-ai-is-trending-now-52a2554b92f8
• Starita, L. (2023, July 6). AI in Marketing: Benefits, Use Cases, and Examples. Persado. Retrieved August 20, 2023, from
https://www.persado.com/articles/ai-marketing/
• Varshini, P. (2020, May 27). Building a Netflix Recommendation System. Medium.com. Retrieved August 20, 2023, from
https://medium.com/analytics-vidhya/building-a-netflix-recommendation-system-7b1fec90f83e
• Whitehouse.gov (n.d.). Blueprint for an AI Bill of Rights. https://www.whitehouse.gov/ostp/ai-bill-of-rights/