Translating AI Election Strategies in India to Business Stakeholder Engagement: Opportunities and Risks

In a recent article from the New York Times, the spotlight was cast on the burgeoning role of artificial intelligence in shaping the upcoming elections in India. As the world’s largest democracy gears up for another electoral showdown, AI tools are being leveraged to create hyper-personalized messages and engage voters on an unprecedented scale.

In particular, the article mentions that these advanced AI tools allow for the creation of video messages and phone communications that can be automatically generated in any of India’s numerous languages, enhancing the accessibility and reach of these messages. Such capabilities not only streamline the outreach process—requiring just a fraction of the time and financial resources typically spent on traditional campaigning—but they also hold the potential to fundamentally transform electoral strategies.

An A.I.-generated version of Prime Minister Narendra Modi that has been shared on WhatsApp shows the possibilities for hyperpersonalized outreach in a country with nearly a billion voters.

However, it is not perfect. Mr. Modi appears to wear two different pairs of glasses. pic.twitter.com/4UgPfRDLkD

— Ahmer Khan (@ahmermkhan) April 18, 2024

Hyperpersonalisation in Stakeholder Outreach

1. AI Avatars and Personalized Communication

• Corporate Context: Just as political campaigns in India utilize AI-generated avatars for personalized voter outreach, corporations can use similar technology to address stakeholders like investors, regulators, and employees. For instance, an AI-generated avatar of a CEO could deliver personalized quarterly earnings updates to investors, tailored to highlight specific interests such as sustainability initiatives or international expansions.

Examples and Application in Business:

• Example: A multinational corporation could deploy an AI-generated avatar of its CFO to interact with stakeholders across different regions, delivering financial updates in various languages and addressing regional market specifics, much like Modi’s avatar did for voters.
• Data Utilization: AI can analyze past engagement data to determine which financial aspects are most pertinent to different stakeholder groups, similar to how political avatars are tailored to address the concerns of specific voter demographics.

2. Technology Advancements and Efficiency

• Efficient Communication: AI significantly reduces the time and resources required for creating and distributing personalized messages in political campaigns. In business, similar efficiencies can be achieved by automating the production of personalized stakeholder communications, such as detailed financial reports or updates on regulatory compliance.

Examples and Application in Business:

• Example: A tech startup could use AI to automate the creation of investor presentations, each personalised to show specific metrics like user growth or revenue per user, based on the historical interests of each investor group.
• Data Handling and Output: An AI system like the one used by The Indian Deepfaker could be programmed to generate thousands of personalized reports daily, ensuring that every stakeholder receives information that’s most relevant to them, presented in a visually engaging and easy-to-understand format.

Enhancing Engagement Through AI-Driven Personalization

• Personalized Avatars for Routine Communications: Regular updates on company performance, upcoming products, or market changes could be delivered through personalized avatars, providing a more engaging and direct form of communication that captures the attention and interest of stakeholders, much like voters are engaged in India.
• Interactive AI Platforms for Stakeholder Queries: Just as AI chatbots in elections can field voter queries on various topics, AI-driven systems in businesses could allow stakeholders to ask questions about the company’s financial health, future plans, or policies and receive instant, personalized responses.

However, as it becomes increasingly easier for executives of both large and small companies to deliver targeted messages to specific stakeholders through selected media channels, the safeguards to prevent misuse remain underdeveloped.

This opens up opportunities for unscrupulous agents, such as short sellers, to exploit these advanced technologies. A deep fake of a CEO mentioning that his company missed its sales target could potentially depress the company’s share price.

Even if such a scam is quickly detected and the stock recovers within seconds, the brief volatility can lead to significant financial gains for those who orchestrate these schemes. Hence, the urgent need for robust regulatory frameworks to keep pace with technological advancements, ensuring that the transformative potential of AI in communications does not become a tool for financial misconduct.

From Novice to Expert: Transform Your AI Engagements with These Proven Frameworks

ChatGPT, with over 180.5 million users, and Gemini, with 142.6 million monthly users, demonstrate the widespread appeal of advanced AI. Yet, despite these impressive figures, many users barely scratch the surface of their potential. Meanwhile, Claude AI, part of Anthropic AI, draws 1.7 million visits monthly, signaling growing interest.

This blog post offers an in-depth look at three essential guides designed to elevate your prompt engineering skills from basic to expert. Master these powerful frameworks and unlock a new level of precision and creativity in your AI interactions.

Writing Effective Prompts with Gemini for Google Workspace

Integrate the power of AI into your everyday tasks with Gemini for Google Workspace. Built on the foundations of real-time collaboration, Gemini enhances your productivity and creativity within familiar applications like Gmail, Google Docs, and Google Slides. This section offers foundational techniques to craft effective, instructional prompts, turning Gemini into your AI-powered assistant. Engage in dynamic conversations with your AI, where each prompt serves as a stepping stone to more efficient workflows.

Key areas for writing effective prompts include:

• Setting clear, instructional goals
• Maintaining a conversational tone
• Being specific with requests
• Iterating based on feedback

While the potential of generative AI is vast, adhering to these best practices can significantly enhance your interactions, ensuring privacy and security without compromising functionality.

Click to access gemini-for-google-workspace-prompting-guide-101.pdf

“The Prompt Guide” by Dave Birss”

• This PDF guide covers the C.R.E.A.T.E. framework for effective prompting, including elements like character, request, examples, and tone.
• It provides practical tips and examples for crafting prompts to get desired outputs from AI models.
• https://davebirss.com/documents/the_prompt_guide.pdf

“Fundamentals of Prompt Engineering” by National Bank of Greece:

• This PDF guide delves into the fundamentals of prompt engineering, covering topics like prompt length, model choice, and target audience.
• It offers considerations and example prompts to help readers understand and apply prompt engineering techniques.
• https://developer.nbg.gr/sites/default/files/PromptEngineeringF.pdf

“Chatbot Prompting: A guide for students, educators, and an AI-augmented workforce”:

• This research paper provides a comprehensive guide on chatbot prompting, including best practices and strategies for different use cases.
• While not a standalone PDF, it offers in-depth insights into the prompt engineering process for chatbots and conversational AI.
• https://www.researchgate.net/publication/367464129_Chatbot_Prompting_A_guide_for_students_educators_and_an_AI-augmented_workforce

Yann LeCun’s Twitter List

Even if you did not know Yann LeCun you might have benefitted from one of his early breakthroughs. He developed an algorithm for recognising handwriting is known as the “LeNet” algorithm. LeNet is a convolutional neural network structure proposed by LeCun et al. in 1998, specifically designed for handwritten digit recognition tasks. This algorithm has been foundational in the field of deep learning and has significantly contributed to advancements in image recognition and machine learning applications.

Yann André LeCun, born on July 8, 1960, in Soisy-sous-Montmorency, France, is a distinguished French-American computer scientist renowned for his pioneering contributions to the fields of machine learning, computer vision, mobile robotics, and computational neuroscience. He holds the prestigious position of Silver Professor at the Courant Institute of Mathematical Sciences, New York University, and serves as the Vice President and Chief AI Scientist at Meta (formerly Facebook).

LeCun’s academic journey began with a Diplôme d’Ingénieur from ESIEE Paris in 1983, followed by a PhD in Computer Science from Université Pierre et Marie Curie in 1987. His doctoral thesis, “Modèles connexionnistes de l’apprentissage” (Connectionist Learning Models), laid the groundwork for his future research in neural networks and machine learning.

In 1988, LeCun joined the Adaptive Systems Research Department at AT&T Bell Laboratories in Holmdel, New Jersey, where he developed several innovative machine learning methods. Among his notable contributions is the development of convolutional neural networks (CNNs), a biologically inspired model of image recognition that has significantly impacted the field of computer vision.

LeCun’s work extends beyond academia to practical applications, including the development of the DjVu image compression technology and the Lush programming language, both co-developed with colleagues Léon Bottou and Patrick Haffner. His contributions to optical character recognition and computer vision using CNNs have been widely recognised and utilised in various industries.

In 2018, LeCun, along with Yoshua Bengio and Geoffrey Hinton, was awarded the Turing Award, often referred to as the “Nobel Prize of Computing,” for their collective work on deep learning. This trio is sometimes affectionately known as the “Godfathers of AI” and “Godfathers of Deep Learning,” highlighting their foundational role in the advancement of artificial intelligence.

Throughout his career, LeCun has been actively involved in the academic community, contributing to numerous publications and serving on the editorial boards of several prestigious journals. He is also a sought-after speaker and educator, sharing his expertise in AI, machine learning, and robotics.As a testament to his influence and contributions to the field, LeCun is an ACM Turing Award Laureate and maintains an active presence on social media platforms like Twitter, where he engages with the broader AI and machine learning community.

Yann LeCun has made several notable contributions to the field of artificial intelligence (AI), particularly in the areas of machine learning, computer vision, and deep learning. His work has significantly shaped the current landscape of AI research and application. Here are some of his key contributions:

1. Convolutional Neural Networks (CNNs): LeCun is one of the pioneers of CNNs, a class of deep neural networks that are particularly effective for analysing visual imagery. CNNs have become a foundational technology in computer vision, enabling a wide range of applications from image and video recognition to medical image analysis12.
2. Optical Character Recognition and Computer Vision: LeCun’s work on optical character recognition (OCR) using CNNs has been groundbreaking. His research has improved the ability of computers to recognise text from images, which has practical applications in various fields such as digitising written documents and automatic number plate recognition1.
3. Deep Learning Techniques: Along with Yoshua Bengio and Geoffrey Hinton, LeCun has been instrumental in the development and popularisation of deep learning techniques. Their collective work on deep neural networks has enabled significant progress in disciplines including speech recognition, natural language understanding, and robotics3.
4. Graph Transformer Networks: LeCun developed the Graph Transformer Networks method, which is similar to conditional random fields, and applied it to handwriting recognition and OCR. This method represents another significant contribution to the field of machine learning, showcasing his ability to innovate and apply AI techniques to practical problems1.
5. DjVu Image Compression Technology: LeCun is one of the main creators of the DjVu image compression technology, along with Léon Bottou and Patrick Haffner. DjVu is an open-source technology that allows for the distribution of high-resolution images in a compressed form, facilitating easier and faster sharing of detailed images online1.
6. Lush Programming Language: In collaboration with Léon Bottou, LeCun co-developed the Lush programming language, which is designed for research in machine learning and other numerical computations. Lush provides a flexible and powerful environment for developing AI models and algorithms1.
7. Advocacy for Open Source AI: LeCun has been a vocal advocate for open source AI, arguing for the sharing of AI technology and models to foster innovation and ensure broader access to AI advancements. His stance on open source AI reflects his commitment to the democratisation of AI technologies5.

Twitter URL	Handle	Description
https://twitter.com/_akhaliq	@_akhaliq	AI research paper tweets, ML
https://twitter.com/Gradio	@Gradio	(acq. by @HuggingFace) dm for promo
https://twitter.com/AIatMeta	@AIatMeta	AI at Meta
https://twitter.com/YiMaTweets	@YiMaTweets	Chair Professor in AI, Director of IDS, Head of CS, HKU; Professor of EECS, Berkeley; Author of Book: High-Dim Data Analysis, http://book-wright-ma.github.io.
https://twitter.com/alfcnz	@alfcnz	Musician, math lover, cook, dancer, , and an ass prof of Computer Science at New York University
https://twitter.com/PyTorch	@PyTorch	Tensors and neural networks in Python with strong hardware acceleration. PyTorch is an open source project at the Linux Foundation. #PyTorchFoundation
https://twitter.com/soumithchintala	@soumithchintala	Cofounded and lead @PyTorch at Meta. Also dabble in robotics at NYU. AI is delicious when it is accessible and open-source.
https://twitter.com/jeremyphoward	@jeremyphoward	Co-founder: @AnswerDotAI & @FastDotAI; Hon Professor: @UQSchoolITEE; Digital Fellow: @Stanford
https://twitter.com/Michael_J_Black	@Michael_J_Black	Director, Max Planck Institute for Intelligent Systems (@MPI_IS). Chief Scientist @meshcapade. Building 3D digital humans using vision, graphics, and learning.
https://twitter.com/yoavgo	@yoavgo	(((ل()(ل() ‘yoav))))
https://twitter.com/CSProfKGD	@CSProfKGD	#CS Associate Prof @YorkUniversity, #ComputerVision Scientist Samsung #AI, @VectorInst Faculty Affiliate, TPAMI AE, #CVPR2024/#ECCV2024 Publicity Co-chair
https://twitter.com/ykilcher	@ykilcher	I make videos. Skill > Destiny. vi / vim
https://twitter.com/giffmana	@giffmana	Researcher (Google DeepMind/Brain in Zürich, ex-RWTH Aachen), Gamer, Hacker, Belgian. Mostly gave up trying mastodon as lb@sigmoid.social
https://twitter.com/huggingface	@huggingface	The AI community building the future. http://hf.co/jobs #BlackLivesMatter #stopasianhate
https://twitter.com/NeurIPSConf	@NeurIPSConf	New Orleans, Dec 10-16, 23. https://neurips.cc/Register/view-registration… Tweets to this account are not monitored. Please send feedback to townhall@neurips.cc.
https://twitter.com/sirbayes	@sirbayes	Research Scientist at Google Brain / Deepmind. Interested in Bayesian Machine Learning.
https://twitter.com/percyliang	@percyliang	Associate Professor in computer science @Stanford @StanfordHAI @StanfordCRFM @StanfordAILab @stanfordnlp | cofounder @togethercompute | Pianist

Sources:

1. LeCun, Y., Bottou, L., Bengio, Y., & Haffner, P. (1998). Gradient-based learning applied to document recognition. Proceedings of the IEEE, 86(11), 2278-2324.
2. LeCun, Y., Boser, B., Denker, J. S., Henderson, D., Howard, R. E., Hubbard, W., & Jackel, L. D. (1989). Backpropagation applied to handwritten zip code recognition. Neural computation, 1(4), 541-551.
3. Bengio, Y., Courville, A., & Vincent, P. (2013). Representation learning: A review and new perspectives. IEEE transactions on pattern analysis and machine intelligence, 35(8), 1798-1828.
4. LeCun, Y., Kavukcuoglu, K., & Farabet, C. (2010). Convolutional networks and applications in vision. Proceedings of the IEEE, 98(8), 1398-1410.
5. LeCun, Y. (2019). Yann LeCun on Twitter: https://twitter.com/ylecun

In Search of Positive Streak, Snapchat Declares “Less Social Media – More Snapchat”

Oxford Circus Tube Station

In a surprising twist in the narrative of digital communication platforms, Snapchat has unveiled a bold advertising campaign: “Less social media. More Snapchat.” This slogan, at first glance, may seem paradoxical for a platform that is inherently a part of the social media landscape. However, a closer examination reveals the layered strategy behind this message.

The State of Social Media Today

The current social media environment is riddled with concerns over privacy, mental health, and content authenticity. Traditional platforms are facing scrutiny for their role in exacerbating these issues, leading to a growing dissatisfaction among users. In response, people are increasingly looking for alternative ways to stay connected that don’t compromise their well-being or personal data.

Snapchat’s Positioning and User Growth

Snapchat appears to be positioning itself as such an alternative. Despite the challenges it faces, including a significant decline in its stock value – nearly 90% from its all-time high – Snapchat has continued to grow its user base. The platform reported a 10% increase in global daily active users, reaching 414 million by the end of 2023. This growth is particularly notable among Generation Z users, with around 63% using Snapchat, positioning it just behind TikTok in popularity.

Financial Performance and Market Challenges

However, Snapchat’s financial performance tells a different story. The platform has struggled with monetization, with average revenue per user (ARPU) decreasing by 5% year over year to $3.29 in the fourth quarter. This is concerning, especially since North American users, who are the most profitable, have shown stagnant growth. The revenue generated from this demographic stands at $900 million, constituting 66% of the total revenue, yet the user base in this region has only seen a marginal increase.

The stock is currently trading at about $11 per share after reaching an all time high of $83.11 in 2021.

The Messaging Behind the Campaign

With this context, the “Less social media, More Snapchat” campaign can be interpreted as Snapchat’s effort to redefine what it means to be a social media platform. The company seems to be distinguishing itself from the traditional, often toxic, social media experience by emphasizing its unique features – ephemeral content, a focus on personal rather than public spaces, and innovative communication tools.

This approach suggests a strategic pivot towards promoting Snapchat as a safer, more private, and more authentic space for communication and connection, particularly appealing to the younger generations disillusioned with conventional social media practices.

Is this the end of a negative streak?

Snapchat’s latest ad campaign could be seen as a bold move to capitalise on the current social media fatigue. By advocating for “More Snapchat,” the platform is not merely seeking to increase its user engagement but is attempting to redefine its market position in a crowded and evolving digital landscape.

The success of this strategy will depend on several factors: user reception to the campaign, Snapchat’s ability to effectively monetise its growing user base without compromising the user experience, and its capacity to innovate in a way that respects user privacy and fosters genuine connections.

What Snapchat cannot control are the moves from current competitors (from Tik Tok to Pinterest) and whether one or more AI based social media might fundamentally change the industry.

Social Media Brands (Partial List)	Ticker Symbols or Ownership
Snapchat	SNAP
Facebook	META (formerly FB)
Twitter	Now owned by Elon Musk
Pinterest	PINS
LinkedIn	MSFT (as it is owned by Microsoft)
TikTok	(TikTok is owned by ByteDance, which is not publicly traded)
Reddit	Reddit is preparing for an initial public offering (IPO) with a target valuation of up to $6.5 billion.

The Battle of Language Models: ChatGPT’s Latest Update in the LLM Arena

OpenAI has introduced some exciting developments that enable users to create custom versions of ChatGPT for specific purposes, known as GPTs. These GPTs can be tailored to be more helpful in daily life, at work, or at home, and you can even share your creations with others.

https://openai.com/blog/introducing-gpts

Meet the GPTs

1. Easily Build Custom GPTs: The process of building your GPT is remarkably accessible. No coding skills are required. It’s as simple as initiating a conversation, providing instructions and additional knowledge, and selecting the functionalities you want your GPT to perform.
2. Use Cases: GPTs have a broad range of potential uses. For instance, they can help with learning the rules of board games, teach math to kids, or even assist in designing stickers. You can create GPTs for your own use, within your company, or make them available to the public.
3. Example GPTs: OpenAI has already provided example GPTs for ChatGPT Plus and Enterprise users, including offerings from Canva and Zapier AI Actions. They plan to make GPTs available to a wider user base soon.
4. The GPT Store: In the near future, OpenAI will launch the GPT Store, where users can publish their GPTs. These creations become searchable and may gain recognition and popularity. The GPT Store will have categories like productivity, education, and entertainment. Users may also have the opportunity to earn money based on the usage of their GPTs.
5. Privacy and Safety: OpenAI is committed to ensuring privacy and safety. Your chats with GPTs are not shared with creators. You can also choose whether data can be sent to third-party APIs. Users have control over whether their interactions with GPTs can be used to improve and train the models.
6. Safety Measures: OpenAI has established systems to review GPTs against their usage policies, aiming to prevent the sharing of harmful content. They are also allowing builders to verify their identity, enhancing trust in the system.
7. AI Agents: OpenAI envisions a future where GPTs can take on real-world tasks, often referred to as “agents.” They plan to move incrementally toward this future, taking into account technical and safety considerations.
8. Connecting GPTs to the Real World: Developers can define custom actions by making APIs available to GPTs. This allows GPTs to integrate external data or interact with the real world, opening up possibilities like connecting to databases, emails, or e-commerce orders.
9. Enterprise Customization: For enterprise users, GPTs offer the ability to create versions tailored to specific use cases, departments, or proprietary datasets. Companies like Amgen, Bain, and Square are already leveraging internal GPTs for various purposes.
10. Community Involvement: OpenAI believes that the most impressive GPTs will come from the community. They aim to involve a wider range of people in deciding what to build and are focused on building safer and better-aligned AI.
11. ChatGPT Plus Update: ChatGPT Plus has been refreshed with information up to April 2023. The model picker has been simplified for a more user-friendly experience, and it now integrates DALL·E, browsing, and data analysis without the need to switch models. Users can also attach files to let ChatGPT search PDFs and other document types.

The European Union Sets the Bar for Consumer Rights in the Digital Age

What is the European Union for? Fundamentally, it exists to improve the lives of European citizens through economic and political cooperation. Over the past few weeks, the EU has come up with another set of practical, consumer-friendly improvements that could become near universal standards. In two landmark initiatives, the EU will substantially strengthen privacy rights, curb harmful online content, and simplify the way people across Europe charge their devices.

The EU’s trailblazing Digital Services Act is a game-changer for protecting privacy in the digital age. As of August 25, 2023, major technology companies like Meta and Google must comply with strict new rules giving users unprecedented control over their data and online experiences. For instance, Europeans can now easily opt-out of personalised advertising based on surveillance of their online activity. This shuts down manipulative filter bubbles. More broadly, the legislation forces Big Tech to be far more transparent about their algorithms and share information with researchers. Fines for non-compliance can be up to 6% of global revenue. By standing up to corporate power, the EU is asserting that citizens’ rights matter more than profits.

In another huge win for consumers, the EU has approved a law requiring all new smartphones, tablets and cameras to use USB-C chargers by 2024. This universal standard will massively simplify gadget charging and allow more mixing and matching of brands. No more piles of proprietary cables and adaptors! Apple is the most affected, since iPhones still use their own Lightning port. But they now have no choice but to adopt USB-C in Europe. This kind of pragmatic, forward-thinking regulation improves daily life for millions.

With the passage of these two impactful laws, the European Union has a prime opportunity to demonstrate its continued relevance and value to citizens across Europe. At a time when skepticism of the EU project simmers in some corners, enacting pro-consumer regulations that tangibly improve people’s digital experiences could help bolster public confidence.

Links:

EU Policies Digital Services Act

The Guardian on the DSA

I Pre-Wrote This Article

In order to convey a sense of anticipation, exclusivity, and novelty while glossing over the price and effort of engaging in purchasing goods and services, a notable trend has emerged: the pervasive use of the prefix “pre.”

Linguists correctly argue that language, any language, is always in flux. The mistakes, misspellings and mispronunciations of yesterday have become today’s standard. Commerce and marketing have also made recent contributions to the English language in some cases, simply by adding a single letter in front of existing words. Thus commerce has expanded into e-commerce and marketing into e-marketing.

The letter and prefix “e” in these cases stands for “electronic.” It is not in contradiction with the word it had been assigned to, but to highlight a specific subset of a general category.

But commerce, with or without the “e,” cannot stand still. Profits need to be chased and maximized. New concepts need to be invented, whether or not they make actual sense.

Unraveling the Temporal Enigma: “Booking Before Booking”

Let’s say you go to your favorite e-commerce website in order to secure accommodations or services for a future date. To be more precise, you have heard that an artist is going to perform in a venue nearby and are eager to buy a ticket for the concert. What verb will you use in order to describe the steps involved in making sure you will be there enjoying the epic ballads of that artist? Will you “book” a ticket? No, you will “pre-book.”

What is the point of the “pre” if the steps involved in “booking” versus “pre-booking” are the same?

One of the ways in which companies increase revenues and profits is to make the act of purchasing an item happening earlier than ever. However, parting with money sooner is not in the best interest of the consumer. On the contrary, the best interest of the consumer is to play as late as possible.

This is how the prefix works its magic. It takes the consumer’s attention away from considering her own financial interests and redirects it towards the idea of having been efficient and in control of her life. In other words, she can tell herself “my life is chaotic, but in this particular instance at least, I have pre-booked, I am way ahead of my timetable.”

Other notable spawns of the “pre-booking” hoax are “pre-paying” and “pre-registering.” Perhaps there is also a sense of being ahead of the masses, who are simply “paying” and “registering.” But I am not the masses since I am using the “pre-payment” option. You won’t catch me slogging it out like a mere mortal.

Why is the use of the prefix “pre” such a great mystification? “Pre” denotes prior action and therefore introduces an intriguing paradox: how can something be done before it is done? How can we pay before paying, or book before booking?

Overlooking this apparent subversion of time is only possible in a society which puts enormous value in the illusion of privileged access to goods or services.

Love Thy Neighbour, But Not Too Much

Car dealerships and thrift shops have also contributed to enriching the English language with the introduction of “pre-owned” and “pre-loved” instead of the straightforward “second-hand.” These adjectives, while intended to evoke sentiments of endearment, inadvertently serve to obscure the pragmatic reality that these goods were once in the possession of others, whose hygiene practices and personal histories remain entirely unfamiliar to prospective buyers.

By introducing the prefix, marketers navigate delicately around the discomfort that may arise from pre-judging the previous ownership by strangers. The result is a carefully crafted narrative that seeks to transform the transactional act of acquiring items once used by total strangers into a sentimental journey of shared affection.

Previewing the Conclusion

The benefits of this subtle but steady linguistic evolution have not spared this writer. I am well aware that your time is precious, just like mine.

So rest assured, you did not spend time reading this article… you gained time. Why? Because I “pre”-wrote this article so you and I would have more time to pre-order the future.

The Difference Jackson Pollock Made

Two almost identical prompts for Midjourney, except I added “Jackson Pollock” in the second one.

the ruins of carthage with carthaginian galleys and a mosq, moon crescent, colours by roy liechtenstein, in the style of zdzislaw beksinski, enki bilal and miyazaki, chiaroscuro by caravaggio :: a mystical lake that is a gateway to another dimension, a lunar clock, in the style of david hockney , joan miro, escher, emc turner, de chirico, –ar 16:9 –q 2

And this is the second attempt with “Jackson Pollock”.

the ruins of carthage with carthaginian warships galleys, moon crescent, colours by roy liechtenstein, in the style of zdzislaw beksinski, enki bilal and miyazaki, chiaroscuro by caravaggio :: a mystical lake that is a gateway to another dimension, in the style of jackson pollock, david hockney, joan miro, escher, emc turner, de chirico, –ar 16:9 –q 2 –v 5

The Science of AI Art Addiction: Dopamine, Novelty, and Midjourney

Prompting My Dopamine Receptors

What kind of image do you think will be generated by this description “a robot listening to music with headphones in a record shop, he holds a record in his hand to closely inspect with x-rays :: esoteric musical notations, scales notes in the background. comic book style, dreamy, zdzislaw beksinski :: A dream world, strange and surreal landscapes, hieroglyphs, psychoanalisys, esoteric mayan symbols, floating objects, surrealism, abstract expressionism, magical realism, René Magritte, Jackson Pollock, Gabriel García Márquez, pieter bruegel, mc escher, oil on canvas,”

And now let’s see if we can get a dopamine rush and see what the prompt generated: https://cdn.midjourney.com/7a46fd09-0fa8-431e-b358-86a46b718773/0_3.webp

Riding the Reward Prediction Error Rollercoaster

The first time I used Midjourney, I spent all the allotted 3.3 hours of CPU time in the course of an evening. I simply could not stop myself trying out new prompts and discovering what the AI might come up with. As we shall see later on, I was like a casino gambler who had lost track of time, because my mind was so actively focused on one particular task. When the Midjourney bot sent me this unequivocal message: “You’ve run out of hours on your plan for this month” my disappointment was real but I was also surprised at my gut reaction. The question I asked myself was “How did I get so addicted so quickly”?

By coïncidence, a few weeks following this episode, I happened to stumble upon the book: “DOPAMINE: The Molecule of More by Lieberman & Long”. As I read about how Dopamine influences our choices, moods and behaviours, the answer to my addiction, disappointment and success of Midjourney became clear.

The principal reason behind the success of Midjourney, is that it provides users with a highly rewarding dopamine rush. This happens because the process of entering a text prompt like the one above and waiting to see what image Midjourney will generate, is filled with anticipation and excitement.

This anticipation triggers a dopamine response, even before the image is revealed. The scientists who studied this phenomenon named it “reward prediction error” – we constantly make predictions about what’s coming next, and when the outcome is better than expected, our brain rewards us with a hit of dopamine.

Much like a rat receiving random food pellet rewards, Midjourney users experience reward prediction errors each time they use the tool, as the images generated are unpredictable and often surpass expectations. The timing of the dopamine rush is optimised too – Midjourney takes around 20 seconds to generate an image after a prompt is entered. This wait time is long enough to build eager anticipation, but short enough to quickly deliver the dopamine hit of seeing the end result.

When that unexpected, amazing image pops up on the screen, it’s a thrilling moment as users take in something entirely new that didn’t exist seconds before. The image is a projection of the AI’s understanding of the user’s words. Seeing how that projection compares to the reality the user imagined is fascinating. There’s an immense satisfaction in having commanded the creation of something novel.

This sense of creation and discovery triggers additional dopamine, the “feel good” neurotransmitter associated with pleasure and motivation. The dopamine hit makes users feel empowered and pushes them to continue experimenting with new prompts and honing their Midjourney skills.

Mastering the AI Muse

Over time, users get better at prompting Midjourney to generate images that better match what they envisioned. This mastery of communicating with an advanced AI system to bring their ideas to life is deeply rewarding. Midjourney has tapped into key elements of human psychology – creation, imagination, mastery – to drive compelling, dopamine-releasing experiences for users. Crucially, what the AI creates will never be exactly like what the user had imagined, although it can be much better. This further adds to the dopamine reward cycle thus making the addiction more likely and stronger.

Finally, there is no limit to the prompts that can be tried: from more practical uses such as icons for your presentations to logos for your coffee, to cute “whimsical” watercolours, to mixing real and imagined creatures using different artistic styles and techniques. If your imagination draws a blank, you can always get inspiration from the myriad creations by fellow users.

Much like a slot machine optimised to dole out random rewards, Midjourney offers users an addictive experience fueled by anticipation and novelty-seeking. By delivering ongoing reward prediction errors with optimised timing, Midjourney has hit upon a dopamine loop which explains its viral success and popularity.

Learn More About Midjourney:

• https://www.midjourney.com/home/ – The official Midjourney website to sign up and learn more
• https://docs.midjourney.com/ – Midjourney documentation and guides
• https://www.midjourney.com/showcase/ – Gallery of Midjourney image creations

Social:

• https://discord.gg/midjourney – Midjourney Discord server
• https://www.reddit.com/r/midjourney/ – Midjourney subreddit
• https://twitter.com/midjourney – Midjourney Twitter account
• https://www.instagram.com/midjourney_/ – Midjourney Instagram account
• https://www.youtube.com/channel/UC-g0c3s6FOew6RqQhKX34ng – Midjourney YouTube channel
• https://www.facebook.com/Midjourney-100900374827025/ – Midjourney Facebook page
• https://www.tiktok.com/@midjourney.app – Midjourney TikTok account

Dopamine and the Brain:

• https://www.healthline.com/health/dopamine-addiction – Overview of dopamine’s role in addiction
• https://www.psychologytoday.com/us/basics/dopamine – Psychology Today’s dopamine guide
• https://www.ncbi.nlm.nih.gov/pmc/articles/PMC2626918/ – The neurobiology of dopamine reward prediction error
• https://moleculeofmore.com/ – THE MOLECULE OF MORE – BY DANIEL Z. LIEBERMAN, MD, AND MICHAEL E. LONG How a Single Chemical in Your Brain Drives Love, Sex, and Creativity – and Will Determine the Fate of the Human Race

Google Announces New Tool To Authenticate Images

In a recent blog post, Google announced it is introducing a new tool called “About this image” to help users assess the context and credibility of images.

With the growing prevalence of misinformation and manipulated images, this tool aims to provide users with important information about an image’s origin, appearance on different platforms, and news articles referencing it.

By clicking on the three dots on an image in Google Images (see gif below) or using Google Lens, users can access this tool. Additionally, AI-generated images will be marked with a label to indicate their origin, and creators and publishers will have the option to add similar labels.

Google continues to develop features to enhance information literacy and empower users to evaluate online content more effectively.

Read more at: https://blog.google/products/search/about-this-image-google-search/