Professors’ picks offers a weekly curated selection of FT articles by and for business school faculty to connect classrooms to current events and to develop students’ critical thinking.
Read all submissions at www.ft.com/bschoolpicks. Save this link in myFT to receive emails alerting you to each new edition. Search the tags for relevant teaching topics. Encourage students to join the debate in the comments section beneath the article.
Comments or contributions? Get in touch at [email protected]
Artificial Intelligence
Why ads are coming to your AI chatbot
Disciplines: Strategy, Marketing, Information Systems, Product Management
Tags: Artificial intelligence, business models, search, AI infrastructure, trust, AI governance, OpenAI
Summary: In 2024, OpenAI’s CEO Sam Altman rejected advertising as a “last resort for us as a business model.” Less than two years later, OpenAI has announced that it will roll out advertising on ChatGPT. OpenAI’s trajectory mirrors a similar about-face by Google’s founders Sergey Brin and Larry Page, who, after being famously dismissive of advertising in 1998, launched Google’s AdWords program in 2000.
A number of economic forces shape the “advertising or not” choice faced by the AI companies. Consumers are increasingly using chatbots as an alternative to search engines. Chatbots are very expensive to train and run. Many industry observers believe that like for other media models, a mix of paid and advertising-supported pricing will probably be the successful consumer-facing business model for AI, and advertising could support free AI-powered services as it did for free search and email. Digital advertising has powered the internet economy for over two decades and now dominates the $1tn global advertising industry. Nevertheless, too much advertising too soon could erode trust in chatbots, and some companies like Anthropic and Cohere remain opposed to advertising and are instead focused on enterprise customers.
Classroom application: The article, which I plan to use in my Executive MBA class, frames a central question about consumer-facing artificial intelligence: how will AI companies eventually monetise their massive capital investments and operating costs? It also sets the stage for discussing the extent to which AI threatens Google and Meta’s highly profitable advertising businesses, and whether concerns about cannibalisation might have shaped their early reticence to roll out AI-centred products like ChatGPT. It provides an excellent basis for discussing how business model choices affect the reliability of information we get from chatbots, how this in turn affects the level of trust consumers place in AI and might “break the core value proposition,” and how a company manages these tradeoffs.
Students are likely already familiar with the unprecedented levels of venture capital that have flowed into OpenAI, Anthropic and others, and with earnings calls in January that revealed how four of the “hyperscalars” (Amazon, Google, Meta and Microsoft) will collectively invest $650bn in AI infrastructure in 2026 alone. They can also draw on well known parallels with the evolution of trust and business models for search and social media.
Questions:
-
How does OpenAI generate revenue today? Is it profitable?
-
Do you see a parallel between OpenAI in 2024-26 and Google in 1998-2000?
-
Should OpenAI embrace the advertising business model aggressively? Lay out the tradeoffs to their CEO.
-
Is the global advertising industry large enough to generate a high enough ROI on the massive AI infrastructure investments you’ve been reading about?
-
What lessons can we learn about digital trust and the future of AI business models from search and social media?
-
Google was widely seen as the leader in AI in the 2010s but was seemingly upstaged by OpenAI in 2022. Do you think Google held back because of cannibalisation concerns? Or was OpenAI’s tech simply better?
Arun Sundararajan, Professor, New York University Stern School of Business
Economics
Europe’s resurgent peripheral economies
Tags: Public debt, fiscal discipline, productivity, labour market reforms
Summary: Following the collapse of Lehman Brothers and the subsequent financial crisis in Europe, the economic performance of Greece, Portugal, Italy and Spain became a source of instability for the Eurozone and the European Union. Some influential voices considered the macroeconomic situation and structural problems of these countries a threat to the European project. The period 2008—2012 was characterised by very high levels of public debt-to-GDP and sovereign bond yields that were unsustainable given expected government revenues. The EU and the IMF decided to provide financial assistance, but under strict conditions of fiscal discipline and a commitment to implement structural reforms, particularly in the labour market.
Years after these difficult measures were implemented, Portugal, Spain and Greece are now among the European Union countries with stronger economic performance, with GDP growth rates above the EU average and above those of large EU economies such as Germany and France. Particularly relevant is the convergence in unemployment levels. Greece and Spain had unemployment rates above 20 per cent in 2012; today they are below 10 per cent.
Labour reforms have made labour markets more flexible and dynamic. Combined with fiscal discipline and financial stability, Spain now attracts immigrants who find jobs and contribute to GDP growth and tax revenues. Spain and Portugal attract talent and investment, although stagnant productivity remains a significant weakness. What was once an EU economic problem has become a reference point for other EU members.
Classroom applications: This article provides an excellent example of how structural changes can alter long-term trends. It includes a very interesting graph showing the evolution of unemployment in a set of European countries. That allows for discussion of unemployment as an economic variable subject to cyclical variability, the idea of convergence, and the types of structural reforms that can impact the economy. The instructor and the students can discuss which changes in labour market regulation are needed to allow for a permanent drop in the level of unemployment, introducing the idea of the natural rate of unemployment and the structural factors that explain the difference observed between countries.
The article also highlights that although employment is growing, productivity is not. It provides a starting point for a debate about the relationship between productivity, employment and GDP per capita, as distinct yet closely linked variables. The instructor can introduce the decomposition of GDP per capita as the product of the ratio of employment over population and productivity (GDP per worker). It can be followed by a debate about the different ways to increase the GDP per capita with an ageing population, the role of technological change and the role of immigration, combined with the relevance of immigrants’ skills and the economic structure of the receiving country.
Questions:
-
What is the relationship between an increase in public debt-to-GDP and government bond yields?
-
Considering the evolution of unemployment across different countries displayed in the graph, is unemployment cyclical? Why?
-
What kind of labour reforms can explain why Southern European countries such as Spain or Greece have converged towards average unemployment levels?
-
What is the relationship between GDP per capita, employment and productivity?
-
What structural factors can explain stagnant productivity?
Pedro Aznar, Associate professor, Esade Business School
AI at work
Is AI making work more intense?
Summary: The FT reports a lively debate on the impact AI is having on people at work. Based on a recent Harvard Business Review analysis, John Burn-Murdoch and Sarah O’Connor develop contrasting opinions. He presents a concerning view of AI’s impact, noting that its use has led to an organic but troubling shift in work habits, with an increase in both the number of hours worked and the intensity of that work. For Burn-Murdoch, this broadening of tasks brings not just unease and exhaustion but also new risks, such as impaired judgment from mental overload. He warns of “cognitive debt,” where AI-assisted projects move too fast for humans to keep up with the finer details.
By contrast, Sarah O’Connor argues the change is not being imposed from the top down by employers, at least for now. Instead, she suggests it is driven by workers themselves, who “just can’t resist doing more” with the tools at their disposal. She points out that professional office work uniquely lends itself to this behaviour, where work expands to fill the time available, making it hard to distinguish between higher output and simply more “busy-work”.
Classroom application: EMBA and MBA students are eager to make sense of how to use AI at work and to understand more deeply not only AI’s impact on efficiency but also on people, including the workers and employees they manage. This article presents contrasting views through a lively exchange between two reporters. I am using this exchange to prompt students to discuss their own experiences of the impact it is having on them and their teams at work.
Questions:
-
What kind of AI tools do you use at work?
-
What impact have these tools had on you or others?
-
To what extent do you feel that AI intensifies your work?
-
Can you remain competitive at work whilst taking some distance from AI?
-
Drawing on the debate, the HBR report, and other readings you have done for today, can you discuss the ethics of AI use at work?
-
What steps do you propose to take going forward in your approach to manage AI implementation?
Additional resources:
AI Doesn’t Reduce Work — It Intensifies It. Harvard Business Review
A multilevel review of artificial intelligence in organizations: Implications for organizational behavior research and practice. Journal of Organizational Behavior
Zahira Jaser, Associate professor, University of Sussex Business School
Negotiations
Energy ministers fail to agree on climate goals as US drives wedge
Tags: Negotiation, Multi-Party, Climate, Energy
Summary: Energy ministers failed to reach agreement on tackling climate change over a two-day International Energy Agency meeting marked by a sustained attack on net zero ambitions by the US energy secretary Chris Wright. Unlike in recent years, ministers did not agree a joint position following the talks, in a sign of the divisions stoked by the US.
Classroom application: This article provides an opportunity to discuss strategies for multi-party negotiations, and in particular, the impact of a spoiler who is trying to derail any agreement.
Questions:
-
Why has the focus of the conversation at this meeting been on ensuring energy security rather than on tackling climate change, as it had been in the past?
-
How can one party to a multilateral negotiation such as this prevent a group of like-minded parties from reaching a consensus?
-
What are the possible risks and benefits to acting as a spoiler and trying to derail the negotiation process in a setting such as this one?
-
What tools or strategies does the spoiler employ to break the unity of the other stakeholders and force its will upon the group?
-
What strategies and skills might parties striving toward an agreement employ to frustrate the spoiler’s attempts at derailing the negotiation?
Moshe Cohen, Master Lecturer, Boston University Questrom School of Business
Got feedback on professors’ picks or willing to contribute? Get in touch at [email protected] or add your selected articles and questions in the comments below.
