The Gen AI Frontier

Shaping a World of Limitless Imagination The Gen AI Frontier: Exploring the Rise, Boundless Discoveries, and Ethical Integration of Generative AI Technologies

Table of Contents

Introduction

Upcoming industry trends in AI

Helping enterprises solve data-related problems with Generative AI

Generative AI: Empowering the Creative Economy through Design

The Necessary Intersection of Generative AI, Responsible AI, and Emotions

In a world disrupted by emerging technologies, Generative AI (GenAI) stands out as a revolutionary force reshaping the landscape of work. Empowering creativity, personalized content, and liberating humans from repetitive tasks, GenAI offers an array of groundbreaking solutions. While concerns about job security linger, enterprise adoption of GenAI unlocks its monumental potential to elevate productivity to unparalleled heights.

This eBook embarks on an uncharted journey into the realm of GenAI, exploring its current advantages, potential hazards, and promising future. As we navigate this unexplored terrain, we must embrace the responsibility to coexist harmoniously with this future-forward technology. Join us in unlocking the transformative power of GenAI and unleashing its full potential. First up, Upcoming Industry Trends highlight the urgency to refine current AI capabilities, particularly in tackling complex tasks. Despite remarkable strides in NLP and computer vision, AI's comprehension of interconnected ideas requires improvement. In this evolving human/AI relationship, human involvement remains vital for fine-tuning AI to address model bias and ethical considerations. Multi-agent learning and generalizable models wield unprecedented disruptive potential, propelling organizations to unparalleled efficiency. Amidst this transformative landscape, Fractal emerges as a trailblazer, actively addressing immediate concerns such as quantum computing, cognitive neuroscience, low-latency models, and ethical AI considerations. Prepare to witness Fractal's forward-looking disruption in various industries. Second, embedded in the dynamic business landscape, enterprises grapple with many data-related roadblocks, from handling big data to privacy issues, cost caps, and prolonged development cycles. Helping Enterprises Solve Data-Related Problems with Generative AI outlines how GenAI, powered by robust Generative Adversarial Networks (GANs), can address these challenges as a game-changer. By tapping into synthetic data, businesses can enhance accuracy, follow privacy regulations, stream- line costs, and accelerate product development.

businesses with numerous use cases. It has expanded its sweep of applications, be it interview preparation, news summarization, songwriting, and code debugging. Howev- er, it has limitations, such as potential inaccuracies and biases. Additionally, and notably, ChatGPT cannot provide real-time information beyond 2021. Generative AI: Empowering the Creative Economy through Design delves into how GenAI bridges skill and resource gaps in the business landscape of the creative economy through a brief walkthrough in understanding the capabilities, shortfalls, and inherent risks to maximize its adoption potential for creative professionals. To close this collection, GenAI's default storage of sensitive data for training raises valid concerns surrounding privacy breaches and unethical practices. The Necessary Intersection of Generative AI, Responsible AI, and Emotions traces the potential, inherent risks, and ethical ramifications of human-GenAI interactions. Stemming from the importance of the anchoring principles of transparency and accountability for responsible AI, the article explores GenAI's inconsistencies, fallacies, and difficulty tracking response lineage that requires safeguards. At the core, reviewing human-GenAI interactions as experiments rather than absolute truth becomes significant in navigating the complexity of GenAI. With the beckoning of the future, a deep understanding of GenAI's positives and pitfalls is fundamental for businesses to steer into the forefront of innovation and sustainable growth. Generative AI enables businesses to optimize workflows, streamline processes, and make better decisions, pushing teams to expand the horizons of innovation and creativity and redefine what is achievable. Nested within a balanced equation of co-existence between technology and humanity, businesses can thereby unlock the true potential of generative AI for the good of humankind as a whole.

Progressing further into the journey, OpenAI's ChatGPT has broken through the hype as a versatile tool, empowering

Upcoming industry trends in AI

Author: Suraj Amonkar | Client Partner | AI@Scale, Machine Vision and Conv AI

Not quite human... yet One of the biggest questions asked regarding where AI is headed is whether its complexity will eventually rival that of the human brain. Our ability to create mind maps to link seemingly unconnected ideas is an intriguing concept that begs the question: can machines develop similar capabili- ties? Although natural language processing and computer vision technologies have advanced significantly, they still lack a fundamental understanding of the interrelatedness of ideas. This is especially evident in scientific knowledge — machines currently can’t understand complex concepts such as gravity and space-time. While some research has explored embedding knowledge graphs into models, the absence of a comprehensive understanding of concepts and their relationships is a significant challenge AI faces. In the immediate and medium-term future, human interven- tion will still be required to ensure that AI produces accurate and appropriate results. The question we should be asking in the longer term is not whether AI will replace humans but how much human intervention will be required for a specific task. As Fractal CEO, Srikanth Velamakanni, puts it: “rather than AI replacing humans, we might see humans using AI replace humans not using AI.” Currently, that seems to be where things are headed, which leaves two distinct challenges: Determining the optimal balance of human-machine interaction Restricting the damage — whether intentional or unintentional — humans can cause through program- ming bias and unethical application of AI technology

Artificial intelligence (AI) has come a long way in the past two decades. However, it wasn’t until Google’s Attention Mechanisms paper was published in 2014 that the wider community began to realize its true potential. Since then, we’ve seen the rise of large language models, and the latest iteration — GPT-4 — is a significant inflection point in AI’s journey. Despite the latest developments, however, there are still several challenges facing AI and its relationship with humans. Addressing these complex issues is likely slower and more difficult than advancing the technology itself. But with the global impact, AI promises to deliver; it’s becoming essential to tackle these challenges with the same vigor we’ve shown thus far in evolving AI. The simple/complex dichotomy The evolution of AI seems to be going in the opposite direction of human evolution. Animals prioritize survival, vision, and speech over more complex tasks, yet AI has already tackled some of the more complex problems and is now moving towards mastering natural language process- ing (NLP) which, in many regards, is less complex. While some primitive capabilities that animals possess may be harder to achieve, we are still progressing toward general- izability, which is the holy grail of AI. With the advent of deep learning, we’ve built more generalized models capable of object detection, sentiment analysis, and topic identification. This generalizability is the breakthrough researchers and developers have been striving for, but there is still a long way to go, as AI can still hallucinate and get simple things wrong (e.g., tests on GPT-3 have proven it responds with both correct and incorrect answers). Current AI needs to be fine-tuned to ensure that the solutions to simple tasks aren’t overwhelmingly complicat- ed. Although the complexity of AI’s world model makes this challenging, with continued effort and advancement, we are slowly inching toward true general AI, with more distinct human-like qualities and abilities.

The bias, ethics, and responsible AI dilemma There are fundamental issues with AI that still need to be resolved, such as biases in the models themselves. To make these issues even more complicated, there are differences in societal biases worldwide. The most extreme forms of bias, such as violence and discrimination, are already being addressed, but usually only after the model has been built, which is not an ideal situation. As a global community, we should be striving to improve AI by incorporating ethical considerations into the AI development process itself.

undertaking, there is a lot of potential for a solution like this in rural healthcare, where access to doctors is limited.

Video analysis is also seeing some interesting developments. Using surveillance as the North Star of vision problems, developers are considering whether it’s possible to link and analyze footage from multiple camera feeds. So, to provide a very simplistic example, if the police need to find out who stole a blue bag while wearing a red shirt, could they search through all the footage to find the perpetrator? While it’s not immediately possible — and ethical issues around privacy still need to be addressed — breakthroughs in this sector are imminent.

It’s challenging to determine how much bias should be removed from a model, and what kind of bias should be addressed. Should a model be completely unbiased, or should it reflect the biases present in society? These are complex questions with no simple answers.

To generalize or specialize — that is the question.

At Fractal, we prioritize responsible and unbiased development. We have established a specialized AI ethics team composed of data scientists, data engineers, lawyers, and designers to ensure ethical considerations are considered when building our models. We have already started incorporating privacy-preserving features into our solutions — especially in the field of machine vision. Imminent disruptions The field of AI is moving quickly, and we are likely to see many process innovations and fundamental breakthroughs in the near future. Life sciences, healthcare, and insurance are particularly exciting areas for these kinds of develop- ments right now. Developers are exploring the possibility of using GPT-like models, such as Bio GPT, designed to break down specialist silos and create a “brain” that knows most medical information. Although this is a huge

Another area of development is generalizable and efficient learning models. There are currently different research directions, one of which is developing small models that can perform specific tasks effectively without needing a large model that does everything. Another avenue is exploring multi-agent learning, where multiple models work together and learn from each other. Finally, progress is also being made in multimodal language models, where multiple modes of communication are incorporated into a single generative model — for example, adding machine vision to GPT-4. The next big breakthrough is potentially multimodality in vision, which is still a work in progress. At a minimum, we are several months away from seeing it.

Fractal: Taking AI into the future One of the key things that Fractal does is approach problem-solving in a structured way using a framework called “solving today’s problem, tomorrow’s problem, and the day after’s problem.” The first category includes tasks such as fine-tuning GPT architectures and vision models and forecasting the next best actions, as these are the immediate problems that need to be addressed to make current models work more efficiently.

Fractal’s NLP and Vision team works on the next generation of AI problems, such as exploring quantum machine learning for tasks like protein folding and cognitive neuroscience. This involves understanding the brain’s behavior and decision-making and has led to forming Fractal’s cognitive neuroscience team. Another area of research is how AI models work in low-latency, low-resource environments, the aim of which is to develop models that are small enough to work on edge devices and in ambient AI environments. Finally, the team also considers embedding AI and ethics into the model rather than treating them as external considerations. While these fundamental problems are not application-specific, the belief is that, if we can find solutions to these challenges, we can create truly disruptive technologies. As AI continues to develop, it’s likely to spur new industries and new ways of using technology. However, it will take time for people to learn how to use AI effectively and for the technology itself to overcome its remaining hurdles. While the jury is still out on the ultimate impact of AI, we have reached a critical inflection point, and the next few years are likely to provide some concrete answers.

One of our main areas of focus is quantum computing, which could lead to an exponential change in computation.

Helping enterprises solve data-related problems with Generative AI

Data drives decision-making in business processes, services, and products that move the global economy forward. However, despite the volume of real-world data available, at the enterprise level, the datasets may need to be more extensive or more openly available for use to derive relevant insights. The business case for synthetic data Enterprises need augmented data to solve analytical use cases where real data is insufficient or unavailable. Existing data is copied or sampled until the data set is large enough to produce results. This method creates oversights that can lead to inaccuracy and data compliance issues. Synthetic data addresses many of these problems. Accuracy Instead of just adding repeated records to a data set, using synthetic data to augment the scarce real data leads to more accuracy in replicating the underlying joint data distributions. This is because synthetic data creates fake/synthetic entities or customers based on the behavior and parameters of real-world data, leading to the generation of more representative data overall. Data privacy Many countries have introduced data protection and privacy legislation. While this does not necessarily prevent a business from using customer data, there are tight controls to ensure it is used only for the reason it was collected. Synthetic data negates this issue: it imitates the properties of the real personally identifiable information (PII) data to ensure accurate analytics, but as the data is not linked to a real-world entity, it can effortlessly be created, shared, and disposed of without compromising data subjects or contravening privacy laws. This makes access to relevant data faster and easier. Cost The acquisition, processing, and analysis of external real-world data are expensive for most enterprises. Once a generative model is in place, the cost of generating new data drops, making it more affordable.

Development cycle Prototypes and innovative ideas require rigorous testing to ensure market relevance. Synthetic data enables the opportunity to test and demo products in various iterations, shorten the development cycle and increase market speed and product adoption/success rates.

GANs: From basic to advanced Generative Adversarial Networks (GANs) are a machine learning framework designed and introduced in 2014. GANs consist of two neural networks, a generator, and a discriminator, which compete against each other in a zero-sum game. The end goal is to train the generator to produce data so accurately that the discriminator can no longer identify it as fake. Once sufficiently trained, the generator can synthesize accurate data at scale.

Initial developments in GANs focused on generating images, but the past few years have seen substantial literature around their ability to rate tabular enterprise data. This has opened the door to diverse applications across many industries.

Fractal: Taking GANs one step further

Of course, not all data is static, but it can also include time-series features. So, the second phase included creating synthetic data from real data features such as daily/ weekly/ monthly number of transactions on a credit card or activity on a website. The temporal element we built in the utility can now simultaneously account for both static customer data and variable monthly data.

Working with enterprise data across multiple domains such as consumer banking, retail, insurance, and telecom, Fractal has a rich experience in customer-centric data. Since most data enterprises collect is tabular, we realized that the need to generate synthetic tabular data outweighs other unstructured data types like images or audio. We divided our approach into four phases and completed the first two. In the first phase, we created a synthetic data generator utility that generates a single data set of static features with no temporal element. This includes anything that does not change over time, such as demographics, gender, or occupation. The output is a synthetic version of static data inputted.

When we started our research, the idea was to address the cost, privacy, and scarcity challenges organizations face with tabular data.

Our Experimental Results Synthesized data must accurately represent actual data to produce relevant insights. We discovered that real data sample sizes as small as 10,000 samples are enough to train our architecture to generate synthetic data sets of up to a million unique, accurate records. Currently, there is no universally accepted single metric to measure accuracy, so we have created a custom metric based on Mutual information data matrices calculated on real and synthetic data. We followed it up with MAE calculation on these matrices to arrive at a single 0 to 1 score, closer to 0, indicating synthetic data is remarkably similar to real data. Along with this custom metric, we also use qualitative metrics such as PCA plots, TSE plots, correlations, and distribution comparisons to see univariate distributions of features in the real data and check that the counterparts in the synthetic data were similar. In both phase 1 and phase 2 iterations, our model outperformed traditional GANs.

Data set: Churn data set- bank attrition data with information about the customer demographics and transaction amounts at a customer level.

Samples: 10,000 samples

Features: MAE values for 20 features across various experiments show improvement in performance from left to right:

Univariate comparison plots:

Customer_Age Distribution Plot Comparison

1.0

Original VGAN

Original WGAN

Original CTGAN

Original Data Generator

0.8

0.6

0.4

0.2

0.0

Customer_Age

Months_on_book Distribution Plot Comparison

1.0

Original VGAN

Original WGAN

Original CTGAN

Original Data Generator

0.8

0.6

0.4

0.2

0.0

Months_on_book

Attrition_Flag Comparison across Models

Dependent_count Comparison across Models

10000

Original VGAN

WGAN CTGAN Data Generator

8000

6000

4000

2000

Attrited Customer

Existing Customer

Fast and sophisticated At their core, GANs are a marriage of two neural networks, making them twice as difficult to train as single neural networks. Once training begins, the correct combinations of hyperparameters are essential to producing high-quali- ty, accurate outputs at the end of the process. Hyperpa- rameters like the number of layers, number of neurons in each layer, learning rate, optimizer, and architectural decisions of using the right GAN variant are critical to tune to the specific real input data provided. As GANs are still in their infancy, we discovered that many data science users might need more technical knowledge of deep-learning algorithms to generate high-quality data. To address this, we consolidated our extensive research on which GAN works best for specific data sets into a utility that guides users. Acting as a “plug-and-play” code accelerator, it is designed to automatically select and use the parameters that will produce the best results based on the original input data. Using these recommended parameters generated synthetic data incredibly quickly and with good similarity to real input data.

For highly technical users who want complete control of the GAN setup, the Fractal utility provides maximum flexibility to adjust every individual parameter of both the generator and the discriminator. The ability of the Fractal model to cater to users across a wide range of technical skill sets makes it one of the most sophisticated GANs capabilities currently available. Future applications of GANs We are entering phase 3 of our development, exploring methods of generating multiple data sets in one go without compromising the inter-relationships between them. Our developments have significantly advanced Fractal’s capability in the Generative AI space. We are optimistic that our existing and prospective clients will benefit tremen - dously from this offering and further solidify Fractal’s position as a leader in AI research and solutions develop- ment in the wider analytics industry.

Generative AI: Empowering the Creative Economy through Design Authors: Shivani Gupta | Engagement Manager (Strategic Center) • Mukundhan Kulur | Senior Design Consultant (Fractal Experience)

Artificial intelligence (AI) has firmly established itself as a transformative force across industries, revolutionizing problem-solving capabilities and expanding its reach into various use cases. The creative and design economy is no exception, as it continues to be shaped and molded by the broad spectrum of AI applications, particularly generative AI. In this era of AI intervention, designers and creators find themselves at a crossroads, with a range of emotions from excitement and curiosity to indifference and resistance. However, embracing the future of the design and creative economy requires acknowledging and adapting to this paradigm shift. Traditional team structures no longer define success; the tools, solutions, and even established processes set professionals apart. Nevertheless, amid this disruptive landscape, concerns emerge. Questions like, "Will AI take my job?" or "Will it make my job easier?" reflect professionals’ anxieties. Fortunately, the answer is not a simple yes or no. This article explores the complex relationship between design practitioners, creative professionals, and Generative AI, shedding light on the transformative potential of this collaboration within the design and creative economy. We also look at OpenAI's ChatGPT as an illustrative example, highlighting the key aspects of this evolving landscape. Expanding the Boundaries of AI Conversations When OpenAI launched ChatGPT, it proved to be a game changer, ladling out the impressive and distinct functional- ities of Generative AI technology at the fingertips of everyone, from professionals to the everyday user. For instance, look at Figure 1 to see how it responded when prodded for a self-reflective response on its functionalities.

is designed to reply according

Figure 1: ChatGPT self-reflexive response

to the programmed data model and dataset and trained to reply according to it. The significance of the AI platform's personality and impact on human interactions has yet to be widely known, but necessary for all users to understand. We reached out to the platform to better understand its functionalities and benefits. We aimed to discover how it could be of value to our business.

Figure 2: Different use cases for ChatGPT

As creatives, we have discovered and deployed ChatGPT beyond its original ambit for unexpected purposes. Our extensive list of use cases includes interview preparation, news summarization, song-writing, code debugging, auto-grading homework, and recipe hunting. In addition, ChatGPT has become the go-to solution for many who prefer it over traditional search engines for obtaining

ChatGPT perceives itself as a sophisticated tool designed to augment our daily activities and take the workload off us. It

intended answers.

time needed to produce results for creators. By tapping into ChatGPT, professionals can elevate their work to a higher standard of refinement, allowing them to position and market themselves more effectively. The availability of AI solutions and assistants is growing exponentially, providing creatives and designers with various innovative tools to increase productivity, streamline the thought process, and improve output quality.

But ChatGPT should not be considered an infallible source despite its popularity and many functionalities. It is important to note that speed should not be mistaken for accuracy, reliability, or credibility. While it performs its designated tasks efficiently, ChatGPT operates with certain limitations and biases that restrict the accuracy of its responses. Further, the free version of ChatGPT offers responses only up to 2021 and is not a live platform.

The rise of Generative AI and the way ahead

This limits it from providing real-time information. We decided to ask ChatGPT to list its gaps:

Post-COVID-19 pandemic, technology adoption has gained significant traction, increasing opportunities for individuals to pursue entrepreneurship, side hustles, and become creators. However, thriving in such a hyper-competitive landscape requires significant effort and resources. There exists a gap in the creator ecosystem regarding skill and time and the availability of proficient individuals. Luckily, Generative-AI has emerged as a game-changer during this period, addressing the skill and resource gap with automation. Its advanced capabilities and unique features give creators an edge they could not achieve otherwise by filling the gap in skills and accessibility of resources with automation. It is both avant-garde and user-friendly, making it both intriguing and accessible and requiring only an email to begin exploring its capabilities.

Figure 3: ChatGPT limitations

To better leverage the functionalities of this technology, we must comprehend the reasons behind the rise in populari- ty of ChatGPT and other AI solutions; it took its course to become a technological phenomenon today, and hence understanding the underlying factors is crucial. Before the COVID-19 pandemic, many individuals held a myopic perspective towards AI and intelligent assistants, primarily because of science fiction's portrayal of AI-spurred global domination or a dystopian, catastrophic end of the world. However, ChatGPT has turned this trend around in a short period. ChatGPT has been launched as an open and free-to-use solution that empowers individuals to achieve enhanced efficiency and productivity in their work. As creatives continue to thrive in today's competitive industry, they can significantly capitalize on the versatile functionalities of ChatGPT as an assistant in creating, modifying, and delivering their content. This innovative platform eliminates the barriers to accessing information and finding skilled individuals while significantly reducing the

Figure 4: The rapid rise in popularity of ChatGPT

What are the main criteria we assess AI assistants on?

FACTOR

CRITERIA

ASSESSMENT OF GPT-4

RATIONALE

They have high responsiveness to all prompts & questions.

Relevance

Medium

They have contextual presence and relevance in interactions; they can pick up where the exchange has been left-off. They do have the ability to recognize limits but do not know when the information is inadequate.

Quality

Accuracy

Medium

Precision/ Correctness

Medium

Higih Low High Low High

Agility Reliability Expanse Consistency Retention

Turnaround speed and time is a few seconds.

The coverage of topics is high.

Scalability

Slight variation in prompts can significantly alter the response.

Provocation

Low

AI assistants have the capability to provide prompts, examples, and quick turnaround for generated text, images, and video.

Speed

High

Creativity

They do not yet provide ‘new’ ideas, pushback, or inspiration.

Medium

Variety

Low

Trust

As they are perceived to be accurate/objective, the believability remains higher than it should. AI assistants come with an existential threat, unawareness about data security, ownership/copyright concerns, and lack of content moderation across platforms.

High

Believability

Low/Medium

Psychological Safety

Safety

Medium

Physical Safety

Low

Data Safety

Closing the Gap: Unraveling the AI User Experience and Expectation Although many people are increasingly eager to adopt AI solutions at the foundational level, a lingering distrust still needs to be addressed regarding its capabilities, trustworthiness, and potential in enterprise business. This is primarily due to the recognition that the AI model can exhibit biases that result in challenging predictions and maintain racist or sexist systems.

Notwithstanding these concerns, most individuals bear an optimistic outlook on AI's future in enterprise systems, extending from routine and autonomous work to complex decision-making processes. [1]

Source: OpenAI Examples:

Figure 5: Process flow of OpenAI

Unlocking Insights: Essential Considerations for Future AI Interactions

It is funny that an AI solution asks if we are human. (See Figure 6) To use an AI app, we need to prove ourselves human. Does this mean AI solutions are not allowed to use AI? 1

Figure 6: Captcha Verification screen on ChatGPT

While it says the goal is to collect feedback and make them safer... (See Figure 7) a. What and where is the feedback being collected, and for what? b. How does it make the system safer? 2 How do we know what is incorrect or misleading? (See Figure 7) a. At a deeper level, how would the user identify discriminatory content / inaccurate statements? 3

Figure 7: Misleading information

Data is being collected and reviewed by AI Trainers. If so, what is being collected and how are they being used? 4

Figure 8: Data collection

An overview of what ChatGPT is capable of and is limited to. It is important to know the full capabilities and limitations to know how to best use this. 5

Figure 9: Landing page of ChatGPT

While the response speeds for the free plan are enough for most use cases, subscription and data bandwidth segment the users into different tiers. This creates a lot more implications regarding priority, experience, ease of use, environmental impact, and other criteria. 6

Figure 10: ChatGPT subscription plans

When Gen AI seems self-assured, will you ever question it? Why treat Gen AI as ‘objective’ and not experimental at this stage? 7

How can we reduce human/researcher bias; what limits (of our own) can we test and rethink with AI? 8

What role should it play for us as designers: enabler, assistant, decision-Ad, counterpart, or advisor? 9

How much power and responsibility are we willing to relinquish for efficiency? 10

Conclusion

The potential impact of generative AI on businesses, regardless of size or technological expertise, is undeniably thrilling. Executives must remain vigilant and cognizant of the emerging risks inherent in this early stage of technology development. With intentional use cases, fostering a culture of learning, a change journey, and establishing guardrails and sandboxed environments for experimentation, and creativity & design could be embedded more efficiently in every part of the enterprise.

References: Cohen. (2021, September 25). Research: Mistrust of AI runs deep, but many remain optimistic for its future. Momentive. Image References: Screenshots sourced from ChatGPT

The Necessary Intersection of Generative AI, Responsible AI, and Emotions

Authors: Sray Agarwal | Principal Consultant (Strategic Center) • Shivani Gupta | Engagement Manager (Strategic Center)

How GenAI has transformed the corporate landscape

In the realm of Responsible AI and Human Emotion, it's crucial to view Generative AI (GenAI) through an ethical lens. The stakes are high, as data leakage and privacy breaches can significantly impact an organization's financial standing and reputation. Let's talk about sensitive customer-centric data used as prompts in GenAI models. It's worrisome that these models automatically store such data for future training without giving users a say. This raises ethical concerns about handling customer data and the need for protective measures to safeguard privacy. But that's not all! We've also noticed some quirks with GenAI models regarding consistency and reliability. Surprisingly, they can respond differently even when faced with the same prompts. And imagine this: they sometimes reject their answers when asked for validation! It's hard to rely on a system that behaves inconsistently, right? Plus, there's the challenge of transparency around answer accuracy and attribution. It takes time to track where a GenAI response comes from, which makes accountability a real puzzle. That's why Generative AI must address the issue of attribution head-on. Users deserve to be able to dig into the answers and understand their sources of influence or reference. By bringing more transparency and accountabil- ity to the table, GenAI can embrace responsible practices and meet the expectations of its users.

Figure 2: GenAI for businesses

As GenAI is being used to create human-like content, it's important to look at it from the dimension of societal impact. A lot of time, fake videos, and images (a.k.a. deep fakes) are generated using GenAI algorithm and are used in conjunction with human generated content. In such a scenario, it is impossible to differentiate between various categories of content viz. human generated, AI generated, hybrid generated and everything in between. Shifting our focus to one of today's most pressing issues—privacy—we recognize the significance it holds. he occurrence of numerous privacy and security breaches serves as a reminder of the challenges we face. Take, for instance, the Samsung incident where a code was compro- mised. However, amidst these concerns, there is an opportunity to implement differential privacy mechanisms for prompts, enhancing data protection. Additionally, empowering employees with comprehensive training on these powerful tools becomes crucial. By embracing these measures, organizations can proactively address privacy concerns and foster a positive environment for the use of advanced technologies. Such incidents call for strict regulations and auditing frameworks, as misuse of GenAI can have an irreversible impact on humans and society. It is no more a hidden fact that generative AI is misused to create fake news and videos, which are used to manipulate public opinion and span propaganda.

Figure 1: Generative AI today [1], [2], [3]

We have witnessed the impact of the Cambridge Analytica fiasco, and GenAI is millions of times more powerful.

Virtual Partner: Ex. An interactive space to brainstorm, to practice conversations/pitches, helps you move at a steady pace/get ‘unstuck’ User Beware: Content that gets added to tools may be used for training the model; look at Terms and Conditions of usage + pricing and plans before adding anything beyond sample text/content Creative/Content Assistant: Ex. Images, text, videos generated almost instantly, to test multiple versions of an artifact with small iterations, curating/editing to a desired size, texture, tone, quick compilation & analysis of notes.

Fractal's toolkit & principles: Safeguarding the promise of Generative AI

Figure 3: RAI Framework (Sourced from Fractal)

Figure 5: Screenshot of the Dashboard tool offered by Hyperwrite

Decision-makers, employees, consumers beware!

If we assume GenAI is objective truth and not an actively learning machine that makes mistakes - we are bound to subject ourselves to inaccuracy, and we’d be doing it with chosen ignorance.

At this time, these are few appropriate roles to think of Generative AI in the workplace:

Task Tracker: Ex. Helps draft scheduling, reminders, managing lists, making a work plan Ex. Ability to draft an email to schedule an interview & adapt it based on details (in this case - virtual vs in-person)

User Beware: Inability to appropriately calculate based on time zones despite specifying

Learning Aid: Ex. Exploring a new domain, creating the learning plan & delivering it to you, curating explanations in a way that’s easy for you to understand User Beware: Information provided can be 1) outdated 2) misleading 3) inaccurate; at this stage, we cannot rely on the authenticity, credibility & sourcing provided by Generative AI tools

Figure 6: Screenshot of Hyperwrite rewriting a portion of this article (self-plug) based on criteria we put in (length, tone, etc.)

User Beware: Ownership and rights over content created by/with GenAI is still highly contested and uncertain

Figure 7: GenAI art [5]

Figure 4: Negative influence of GenAI [4]

Perception altered: Examining the mental model when using generative AI Treating my interactions with it as I would with Research When researching, exploration is contained to the hypothesis, the focus is predominantly learning or proving/disproving the hypothesis (ex., AI is good vs AI is bad); it can lead to extreme conclusions and not leave room for quick updating. Treating my interactions with it as I would with digital Products/Services When using products, there is an expectation of objectivity, front-end finesse, and back-end data management, which may not exist for most GenAI products today. With products/services, there is also a clear understanding of who benefits/ - profits from a user’s engagement, but in this case, that isn’t always clear, or relatedly, nor is liability or consequence of ‘bad’ experiences. Recommended: Treating my interactions with it as Experiments When experimenting, there is a playful yet scientific approach to discovery; finding applications, testing limits, pushing boundaries; but with caution.

Conclusion A good metaphor to think of Generative AI in terms of safety is to consider it as a powerful tool in a craftsperson's workshop.

metaphoric chainsaw (replacing people with machines).

Much like a skilled craftsperson, a user should approach Generative AI with caution, understanding its capabilities and limitations, to use it more effectively, safely, and ethically (they don’t have to be in contradiction). Another reason this metaphor works is that when you’re a craftsperson, you’re either thinking trial-and-error/experi - menting with materials/tools, and when you’re building something, you’re thinking about making things to last and things that will stand the test of time. When using GenAI, tapping into either one of these mindsets intentionally will help approach it better; with less errors, less existential threat, and reduced safety risks. Depending on where an organization chooses to play on this matrix - it will require related tools, change journeys, implications & cautions.

Let’s break that metaphor down a bit:

Powerful: Capable of more than you think at first glance; potential to create impact (both positive & harm).

Tool: It can be considered as a suite of tools that can be a facilitator or enabler for your work.

Craftsperson: While anyone can take up crafts, to be a craftsperson is a skill to hone; to identify the right tools for different tasks, to understand effort, preparation, and maintenance of the tools, the space, and oneself. Workshop: You can technically enter a workshop without safety gear and rules; but you will likely cause yourself harm and damage the space if you do; learn and consider safety as you engage with something as small as the metaphoric pin (typing a question into ChatGPT) or the

Sustainable integration of Generative AI in the workplace

EMOTION-INTEGRATED AI & CHANGE MANAGEMENT

goal aligned for users & consumers

managing threat & vulnerabilities

intuitive to people

plots, tracking

privacy, security

RESPONSIBLE AI & GOVERNANCE

de-biasing

Figure 8: Core foundations for sustained and successful implementation; going beyond what is technically feasible towards what is human & business aligned (Sourced from Fractal)

References:

Warren. (2023, February 18). Microsoft limits Bing chat to five replies to stop the AI from getting real weird. The Verge. https://www.theverge.com/2023/2/17/23604906/micro - soft-bing-ai-chat-limits-conversations

Gresseth. (2023, January 26). ChatGPT: Plagiarism super-tool for students or AI brainstorming generator? KSLNewsRadio. https://kslnewsradio.com/1982846/chatgpt-plagiarism-su - per-tool-for-students-or-ai-brainstorming-generator/

Harrington. (2023, May 29). Companies Using AI-Generated Images for Diversity, Instead of Real Models, Are Missing the Point at Best. The Mary Sue. https://www.themarysue.com/compa - nies-using-ai-generated-images-instead-of-models-for-diversity/

Rubenstein. (2023, April 8). Can AI defame you? Morning Brew. https://www.morningbrew.com/daily/stories/2023/04/07/can-ai-defame-you

Quach. (2023, March 16). AI-generated art can be copyrighted, say US officials – with a catch. The Register. https://www.theregister.com/2023/03/16/ai_art_copyright_us - co/#:~:text=US%20law%20states%20that%20intellectual,their%20outputs%20are%20not%20copyrightable

Enable better decisions with Fractal 500® companies. Fractal's vision is to power every human decision in the enterprise, and bring AI, engineering, and design to help the world's most admired companies. Fractal's businesses include Crux Intelligence (AI driven business intelligence), Eugenie.ai (AI for sustainability), Asper.ai (AI for revenue growth management) and Senseforth.ai (conversational AI for sales and customer service). Fractal incubated Qure.ai, a leading player in healthcare AI for detecting Tuberculosis and Lung cancer. Fractal currently has 4000+ employees across 16 global locations, including the United States, UK, Ukraine, India, Singapore, and Australia. Fractal has been recognized as 'Great Workplace' and 'India's Best Workplaces for Women' in the top 100 (large) category by The Great Place to Work® Institute; featured as a leader in Customer Analytics Service Providers Wave™ 2021, Computer Vision Consultancies Wave™ 2020 & Specialized Insights Service Providers Wave™ 2020 by Forrester Research Inc., a leader in Analytics & AI Services Specialists Peak Matrix 2022 by Everest Group and recognized as an 'Honorable Vendor' in 2022 Magic Quadrant™ for data & analytics by Gartner Inc. For more information, visit fractal.ai

Corporate Headquarters

Other Locations

Australia, Canada, China, India, Singapore, UK, Ukraine

Get in touch

Page 1 Page 2 Page 3 Page 4 Page 5 Page 6 Page 7 Page 8 Page 9 Page 10 Page 11 Page 12 Page 13 Page 14 Page 15 Page 16 Page 17 Page 18 Page 19 Page 20

Made with FlippingBook - PDF hosting