Carmem Domingues

This talk explores what are some of the ways that AI is (re)shaping society, and what are its
consequences, focusing on how AI can Empower or Erode Human Behavior.

As AI becomes ubiquitous, and touches ever more facets of our lives, we increasingly defer
to machines to carry out tasks previously done by humans. How, and why is that happening? Are we giving up what makes us inherently humans, and differentiates us from
machines? What makes us uniquely humans? What should we allow machines to do, versus what should we protect as our uniquely human characteristics and abilities?

What are the impacts to the future of labor, human relations (with ourselves and others), global politics and society at large?

This talk will give the audience an understanding of, and make them think about:

• How AI impacts a plethora of aspects of society: from environmental issues, to
  institutional organization, geopolitics, social relations and even human cognition
• What makes us uniquely humans – How AI can both empower and erode human behavior
• How our humanity is impacted by our increased use of AI
• How AI affects from our relationships with ourselves, to the global balance of power
• Ultimately, what should we be delegating to AI systems, and which skills and abilities we should protect as essential to our humanity, and our future
• How we can adapt our own use of AI systems to minimize negative impacts, and maximize positive ones

Especially since the launch of ChatGPT, there has been a lot of confusion about what AI is,
what it is not, and what it can/cannot do. The narratives that surround AI often portray it as
a simple, objective tool, much like a hammer or a calculator. In fact, AI is much more than that. It is a full-blown socio-technical system that is both influenced by, and influences, society.

This talk explores the evolution of AI from its early days to current times and reviews the whole pipeline of AI development, dismantling hypes and myths, and highlighting the true powers of AI. It explains (in lay terms) the technical aspects of how modern AI works, with an eye to helping the audience understand what in the AI development process is done by machines versus by humans. It explores some of the ways in which biases such as human bias or data bias affect the AI development process, and thus the outputs of AI systems.

Ultimately, this talk explores how AI is developed, how it is shaped by societal values, and how it, in turn, shapes society. Discussion includes issues of power at the individual, societal, organizational, institutional, and governmental levels.

In this talk, the audience (even non-technical folks) will learn:

• What is AI
• The true powers and limitations of AI
• How AI is developed
• How different aspects of the AI development cycle influence its output
• How to evaluate AI output and thus better apply it in their day-to-day
• How AI should be adopted
• How AI influences society
• How AI is influenced by society
• How AI is a truly socio-technical system, and what are the implications of this
• How economic, organizational, political and social forces and powers influence both what and how technology is developed, and how different parts of society are
  impacted differently by those decisions
• To cut through unfounded hype and empty promises, and more appropriately
  leverage AI in their lives and organizations to produce positive outcomes and impact

It is not news that tech companies collect data on their users, and sell it for a profit, often without users’ awareness or explicit (vs. coerced) consent. Yet, we continue to increase our use of these tools.

This talk explores some of the hidden ways in which we end up giving a lot more information and power to tech companies than we are aware, the power asymmetries that it creates between users, companies, and even governments, and what are the consequences of that at both the individual and societal level, including impacts to
democracy and global geopolitics.

Audience members will:

• Learn how different technical choices regarding AI development and use across organizations can influence power dynamics between different stakeholders
• Learn about the different AI-related power dynamics at play between users, companies and governments
• Gain insights into how these power dynamics impact each of us at individual, institutional, societal, and global levels, including their significant consequences
  for democracy and geopolitics
• Learn about the role of AI is shifting the global balance of power, and the global order

Research shows that AI, especially Generative AI systems like Large Language Models (think ChatGPT, Claude, Gemini, etc) can influence our thoughts and behaviors. Yet, we increasingly rely on AI chatbots and AI agents to help us with everyday tasks: from the mundane to the complex ones – from GPS navigation to scheduling assistants, computer code development, research and writing help (from text messages to full reports), etc.

Some of these tools can be extremely helpful with cognitive offloading to free us up to focus on the things that really matter to us. However, research increasingly shows that they also impact our future (and even immediate) cognitive abilities, with potentially serious implications to Critical Thinking, and shape our world views, and behaviors, even through
brief interactions.

This talk discusses this research in lay terms, exploring how and why this happens, including the mechanisms through which cognitive erosion and shaping occur. It highlights the implications of this at the individual and societal level, and closes with takeaways for what we can do to protect our cognitive ability, and ourselves.

The audience will:

• Understand how increased use of different types of AI tools can impact cognitive ability and critical thinking, and shape human behavior
• Understand and be able to identify how the different AI systems they use may influence their thoughts and behaviors, and what are its consequences
• Learn how to properly use AI in a way that augments, rather than erodes, human cognition and critical thinking
• Be equipped to not only adjust their uses in ways that better align with their individual goals, but also to develop policies and implementation projects for their
  organizations that can lead to better results

30% of adults report having had some sort of intimate encounter with an AI system, and as many as 20% of teenagers also report having or knowing someone who has had a relationship with an AI.

More and more people are marrying their AI Companions. As AI systems become bigger parts of our lives, more people are resorting to them to fulfill their
emotional needs.

However, this is not always intentional. This talk explores some of the reasons and ways people enter into emotional relationships with AI (this will surprise you!), and what are some of the benefits as well as risks associated with emotional engagements with AI systems.

The audience will learn about how this evolving trend of AI Companionship:

• The surprising and unexpected ways a majority of users enter into emotional relationships with AI systems
• How human-AI relationships develop
• What are the potential benefits and risks of this relationship structure that is becoming ever more common
• How AI systems are often designed to exploit human psychological tendencies that can lead to inadvertent emotional attachment

Research shows that despite all the promises of increased productivity from AI use, 95% of AI Pilots across organizations are being abandoned because adoption is not being worth the investment.

This is not surprising to me – one of the things I have most heard from clients throughout my career is “we want to use AI in our organization.” However, more often than not, that statement is coupled with an inadequate understanding of the capabilities and limitations of AI and what successful implementation requires; and with an approach of “throw AI at the wall and see what sticks,” rather than a clear problem that AI can solve, and a strategy
for how to do it.

This talk explores some of the reasons for this mismatch between expectations and results, and discusses how organizations should think about, and implement AI systems, the right way. It also covers the basics of designing an effective AI Strategy.

After this talk, the audience will be able to start thinking strategically about AI. They will:

• Get an overview of what AI is, what it can/cannot do, and how AI systems work
• Understand why a majority of AI projects are failing and how to prevent it
• Learn the technical and organizational requirements for a successful AI deployment
• Be able to sift through the AI hype and separate fact from fiction, becoming better able to differentiate between empty promises and realistic expectations for AI systems
• Be better equipped to think through build-vs-buy evaluations and how to properly develop and implement AI solutions across diverse organizations and use cases
• Better evaluate what types of organizational problems are good candidates for AI solutions (and which ones are not)
• Learn what types of AI solutions might best fit different use cases
• Learn how to think through whether AI is the right solution for their organization or use case

Much of the debate about AI frames AI as a purely objective tool, without regards to how the ways in which it was developed, and how it is used, can have deep ethical and moral implications.

On the other hand, much of the responsible AI debate tends to focus on issues of bias and metrics to identify and measure its impact, or how to comply with different AI regulatory frameworks. Whereas these are absolutely important, it is crucial that we take a more holistic approach to AI.

This talk helps the audience understand the AI decision, development and implementation
process as a whole, and how to consider ethical implications throughout it, so that they
can develop and deploy AI systems in a more holistically responsible and ethical manner, within proper governance frameworks.

The audience will learn about:

• Issues of bias (including data bias, model bias, developer bias and organizational bias)
• Ethical considerations when deciding whether AI is the best moral and ethically responsible solution
• What limitations it poses throughout the AI development and implementation cycle
• How to address AI limitations (technically or otherwise) ethically and responsibly
• Governance issues, including a panoramic understanding of the major pieces of regulation that impact the governance of AI

Another one of the questions I get asked the most, is how can people best prepare for the future, given how the recent advancements in AI seem to be getting faster, and better, every
day.

This talk takes a historical view of technical (r)evolutions to contextualize the recent
developments in AI. It then discusses in more depth certain recent developments to dismantle some myths and common misunderstandings, and help the audience understand what is actually likely to happen, and how best to prepare for the coming
changes.

In this talk, the audience will:

• Dismantle AI myths and unfounded AI hype
• Gain a solid and realistic understanding of what types of capabilities AI currently has and is likely to develop in the future
• Learn how those are likely to impact different types of work, the labor force, and the future of work as a whole
• Walk away with more clarity about how to prepare themselves and their organizations for the future, including what skills they should consider acquiring or further developing, so they are not left behind