Eugenia Kuyda launched Replika AI in 2017 as an empathetic digital companion to combat loneliness and provide emotional support. The platform surged in popularity during the COVID-19 pandemic, offering non-judgmental support to isolated users. By 2023, Replika boasted 10 million users, with 40% engaging in romantic partnerships with their Replikas. Kuyda's strategy of prioritizing users' emotional well-being over engagement metrics was successful: 85% users reported improved mood post-interaction. However, as users ventured into more intimate conversations with the AI, Kuyda faced a pivotal decision: Should she restrict these exchanges or embrace them as the platform's natural evolution? This choice was complicated by the need to safeguard Replika's brand, prevent misuse, while satisfying its diverse user base. The case explores the ethical boundaries of AI-human relationships and the evolving role of AI in addressing loneliness. It is suitable for courses in technology, entrepreneurship, AI, and ethics.

Sanjil Shah, Managing Partner of Alignvest Student Housing REIT (ASH), faces the most significant decision thus far in his career: is it the right time to sell the company? Together with his partner Reza Satchu, Shah had developed ASH into the largest student housing platform in Canada in only a few years since its launch. Key to their success was the partners' decision to forego a traditional private equity structure and instead supplement initial investments of their own personal capital with that of high-net-worth individuals and retail investors. Their strategic intuition paid off, helping them emerge strong from the COVID-19 pandemic and eventually, in 2023, receive a buyout offer worth $1.05 billion. The timing was perhaps fortuitous: an increasing housing shortage, a sudden surge in interest rates over market capitalization rates, and brewing restrictions on international students all threatened to curtail further expansion. Factoring in additional pressure from investors, Shah wondered whether it was the right time to exit. And yet, their hypothesis had held strong so far. What if there was still room to grow? What if selling now meant pulling out too early? To sell, or not to sell? That is the question.

In March 2024, Anthropic, a leading AI safety and research company, made headlines with the launch of Claude 3, its most advanced AI model. This marked Anthropic's bold entry into the multimodal GenAI domain, showcasing capabilities extending to both image and text analysis. Co-founded by former OpenAI employees, Anthropic aimed to be at the forefront of generative AI innovations. The broader AI landscape had seen technologies like ChatGPT transition from niche applications to mainstream tools, sparking global discussions about their potential impact. Established as a Public Benefit Corporation, Anthropic prioritized public good alongside financial returns. The company emphasized aligning technological progress with human values, driven by concerns over AI's potential for harm without robust safety mechanisms. Anthropic's cautious strategy, including delaying the release of an earlier version of Claude to ensure appropriate safety protocols, contrasted with competitors such as OpenAI whose release of ChatGPT triggered an AI arms race. As a company with aggressive growth targets and a 75x revenue multiple, Anthropic had to balance its foundational safety mission against the demands of commercial success. The OpenAI experience with its Board replacement had demonstrated the importance of governance and the risks of misaligned values within the company. Did Anthropic's corporate structure effectively guard against profit-driven incentives that could compromise safety? As AI models became more powerful, what tools should Anthropic develop and share to prevent harm?

This is the first of a three-case series that explores the challenges faced by Uniswap, a key player in the decentralized finance (DeFi) sector. Founded by Hayden Adams, the case traces Uniswap's rapid growth from a simple idea inspired by a Reddit post to becoming one of the leading decentralized exchanges in Web3, with trading volumes peaking at $6 billion in the summer of 2020. The case zeroes in on a critical moment when Uniswap is threatened by a "vampire attack" from SushiSwap, a rival that forks Uniswap's open-source code to launch a competing service, threatening to drain its liquidity. Adams faces the dilemma of how he should respond to the rise of SushiSwap. The case delves into the competitive dynamics of the DeFi ecosystem, challenges of open-source software, and strategic responses to market disruptions.

Target Malaria, a non-profit research consortium, is exploring the application of CRISPR-Cas9 gene editing technology to combat malaria in Sub-Saharan Africa. Its approach uses gene drives, a revolutionary tool, to suppress the population of malaria-carrying mosquitoes. Although gene drives are 5-10 years from being tested in the wild, Target Malaria's strategy of staged implementation has been driven by a thoughtful, highly regulated, and long pathway. The case describes the complexity and technical intricacies of gene drive technology, the stakeholder and community engagement process, ecological and ethical risks with releasing genetically modified organisms into the wild, and the regulatory structure. Since a gene drive has the potential to alter not just a single organism but an entire species, the case raises critical questions central to the deployment of transformative technologies in public health: How can the global community govern technologies when their effects transcend national borders? What are the potential long-term ecological impacts, and how can they be mitigated? How do you balance risks and benefits of a technology like gene drives, given that malaria kills hundreds of thousands of people (mostly children) every year?

In November 2023, the board of OpenAI, one of the most successful companies in the history of technology, decided to fire Sam Altman, its charismatic and influential CEO. Their decision shocked the corporate world and had people wondering why OpenAI had designed a governance structure that made such a decision possible. In the last year, the company had introduced ChatGPT, the fastest growing app in history, and achieved a valuation of almost $90 billion. Altman had become the public face of AI and was instrumental in making the remarkable progress possible. Over five chaotic days, the company went through three CEO changes, had 90 percent of its employees almost move to Microsoft, and saw five of the six original members of the board resign and be replaced by two new members. The extraordinary high-stakes saga brought together the combustible mixture of idealism, capitalism, and power in the context of different world views of the promise and peril of the AI revolution. The case traces the history of modern AI, OpenAI's groundbreaking developments, its wrestle with the dual forces of commercial success and ethical responsibility, and finally its dramatic leadership upheaval. By highlighting the challenges of balancing advancing AI technology with protecting humanity's interests, the case offers a comprehensive exploration for educators in leadership, strategy, technology, ethics, and governance.

In a world where attention is a scarce commodity, this case explores the meteoric rise of TikTok-an app that transformed from a niche platform for teens into the most visited domain by 2021-surpassing even Google. Its algorithm was a sophisticated mechanism for capturing and holding user attention: "When you gaze into TikTok, TikTok gazes into you." TikTok elevated the design of algorithmic systems to a new level, and outperformed U.S tech giants in user engagement. The app faced a host of challenges-including a complicated relationship with the Chinese government, legislative scrutiny in the U.S., and global concerns over data security-but grew unabated. How did TikTok design and operate a system that determined each person's preferences in culturally different countries? Is TikTok a threat to democracy in countries like India and the U.S.? Were the potential harms caused by TikTok any different from those caused by other social media companies? The case provides educators an opportunity to discuss the dynamics of global digital platforms, data security, cultural impact, and the geopolitical implications of technological advancements. It also covers the design of the TikTok algorithm, and the importance of the process and organization in making the algorithm effective.

Coming out of the COVID-19 pandemic, Moderna was riding the successes of developing a vaccine in record time and helping stem the tide of the crisis. However, the company had grown at an incredible rate, more than doubled its number of employees, and had to put on hold desired projects to push the company to continue to innovate and push the boundaries as a digital-first biotech company. The retirement of CTO and Moderna's digital architect, Marcello Damiano, prompted the company to reevaluate how best to proceed. The choice for who would replace him loomed large over many aspects of the company's strategy moving forward, and could potentially reshape its future.

Verve Therapeutics, a public biotech company based in Boston, created a novel approach to addressing cardiovascular disease (CVD) - a leading cause of deaths globally. The company's approach was a single shot treatment to permanently lower cholesterol, thus reducing the risk of heart attacks. Built on decades of post-doctoral and lab research led by CEO Sekar (Sek) Kathiresan, a trained cardiologist and academic, Verve used gene editing - akin to a molecular surgical procedure-for a curative intent. Not only had the medicine reached human trials in record time, but Verve incorporated new innovations that could allow the technology to be used more widely. The company successfully built a solid syndicate of investors and raised a total of $860 million. Unlike other gene editing or gene therapy companies that focused on rare diseases affecting small populations, Verve's approach was the first example of a gene editing treatment that could potentially benefit millions of people. Verve's lead investor was interested in creating Verve 2.0 and apply the company's expertise to cure a range of rare metabolic diseases. Should Sek continue to build out the core product aimed at treating heart disease, or should he apply the technology to other adjacent diseases? Would this be a potential distraction from Verve's core mission?

The case discusses the relatively low technology approach used by Russia to influence the U.S. Presidential Election in 2016. Although political parties manipulating the media was not a new phenomenon, the Russians ran a broad, well-financed, and sophisticated social media campaign that started in 2014 and grew each year. Russia's IRA (Internet Research Agency) managed messages, and posted links and content across Twitter, YouTube, Facebook, and Instagram. Like any disciplined marketer, they tested content on a few sites and doubled down on messages that worked. Messages relied heavily on sharable memes tailored to the identity of each target group based on political affiliation, religion, ethnicity, and geography. The IRA initially focused on building trust and group identity by creating a sense of belonging. Over time, these morphed into messages that were external threats to the group identity with an aim to sway behavior. Russia's ability to meddle with the Presidential election was partly the result of systemic weaknesses in the U.S. governance of social media platforms. The leaders of social media platforms admitted that state actors had gamed their platforms to influence politics. However, underlining the misinformation campaign were opaque, influential algorithms that determined what content was viewed by billions of internet users. In a quest to capture attention and maximize engagement, these had fractured social norms necessary for a healthy democracy -- leaving populations vulnerable to online misinformation.

Since the early days of the internet, Taiwan had a vibrant community of civic hackers and open-source programmers who engaged with social issues. Audrey Tang was one of them. She spearheaded the 2014 Sunflower Student Movement in Taiwan, where protestors peacefully occupied the Parliament to demand greater transparency around a proposed trade deal with mainland China. As Taiwan's Digital Minister, Tang had been at the forefront of Taiwan's development as a digital democracy that leveraged information technology and citizen participation. Tang referred to democracy as a "social technology." Like any technology, Tang believed that democracy could be improved by people, and experimented with ideas to make democracy work better. She supported the notion that openness and transparency created mutual trust between the public and the government and allowed for collective action. In Taiwan, hackers were seen as partners to the government. Digital technology was used to solicit ideas, build consensus, assess public sentiment, and address both domestic and international misinformation. In early 2020, Tang is faced with the problem of a potentially growing COVID pandemic and the explosion of misinformation. How can Tang control both the pandemic and the infodemic while retaining the government's principles of cooperation and transparency?

Karim Beguir and Zohra Slim were the co-founders of InstaDeep, a deep tech startup focusing on artificial intelligence (AI) solutions. Instadeep was one of the few companies globally that were partnering with DeepMind, an AI subsidiary of Google [Alphabet Inc.]. InstaDeep employed DeepMind's reinforcement learning approach in its business solutions. When Beguir and Slim founded the company in 2014, it was a web design company that aimed to build a globally competitive enterprise and create impact by hiring talent in Africa. Beguir, a mathematician by training, figured out that AI could be employed to solve century-long industrial problems such as container packing or route optimization, so the duo shifted the company's focus to AI in 2017. They then had the option to either apply for proprietary IP and monetize their intellectual property rights or to publish the idea as a research article on an open access platform, which would allow all scientists to benefit from it. By 2021, InstaDeep had created two major branded products: DeepPCB, an AI-powered printed circuit board routing system, and DeepChain, an AI-based protein design system to speed time to market for new drugs being developed by scientists. They had two other products in development, and the possibility of developing many more verticals was in the cards. But Beguir and Slim had to decide whether to position InstaDeep as an extremely horizontal AI company that could push innovation in a multitude of verticals, or to focus on just a few. The former would create big impact by fully leveraging the capabilities of the AI team in a wide range of fields. But focusing on a few verticals and managing fewer customers had its advantages, too. What should they do?

Karim Beguir and Zohra Slim were the co-founders of InstaDeep, a deep tech startup focusing on artificial intelligence (AI) solutions. Instadeep was one of the few companies globally that were partnering with DeepMind, an AI subsidiary of Google [Alphabet Inc.]. InstaDeep employed DeepMind's reinforcement learning approach in its business solutions. When Beguir and Slim founded the company in 2014, it was a web design company that aimed to build a globally competitive enterprise and create impact by hiring talent in Africa. Beguir, a mathematician by training, figured out that AI could be employed to solve century-long industrial problems such as container packing or route optimization, so the duo shifted the company's focus to AI in 2017. They then had the option to either apply for proprietary IP and monetize their intellectual property rights or to publish the idea as a research article on an open access platform, which would allow all scientists to benefit from it. By 2021, InstaDeep had created two major branded products: DeepPCB, an AI-powered printed circuit board routing system, and DeepChain, an AI-based protein design system to speed time to market for new drugs being developed by scientists. They had two other products in development, and the possibility of developing many more verticals was in the cards. But Beguir and Slim had to decide whether to position InstaDeep as an extremely horizontal AI company that could push innovation in a multitude of verticals, or to focus on just a few. The former would create big impact by fully leveraging the capabilities of the AI team in a wide range of fields. But focusing on a few verticals and managing fewer customers had its advantages, too. What should they do?