Stephanie Kim/ProMarket
Three artificial intelligence experts at the University of Chicago reveal the AI trends and topics they will be watching for in 2024.
Anjali Adukia, assistant professor and director of the Messages, identity and inclusion in education (MiiE lab)
Much of the focus on artificial intelligence (AI) in education has centered on writing tools like ChatGPT, a large language model that, among other things, allows users to “converse” with the software to produce a document that meets length requirements and format corresponds to , style and level of detail. Likewise, other AI systems can create images and videos based on prompts. Needless to say, these tools have raised concerns among educators. In either case, it's easy to imagine students secretly using these tools to complete assignments, adding another layer of scrutiny to already overburdened teachers.
However, it's not just for one-off tasks that AI has the potential to fundamentally impact the way students learn and ultimately what they learn. In fact, AI will almost certainly become an everyday part of education, which of course raises concerns and important questions: Over time, as they become more reliant on AI, how will students learn to write? And more fundamentally: Do they even have to learn to write or think for themselves? Or are writing and critical thinking skills only important for creating better AI prompts? Additionally, will AI tools take over art and creativity, from visual arts to music?
For me, introducing AI into teaching is more than just theoretical. In my courses, I have adapted assignments to incorporate AI while also asking students to challenge the resulting text by checking for errors, suggesting improvements, and generally critiquing the text for content, style, and effective message. In order for students to successfully complete this task, they must understand the material, think critically, and ultimately be able to express their thoughts clearly in writing and orally. AI can certainly be a useful tool: it can help create hypotheses, suggest ways to improve writing, and summarize texts. So, if used with care and caution, AI can complement education while improving our learning and workplaces, raising the bar for productivity and efficiency while leaving time to focus on more creative tasks.
However, my relative optimism is greatly tempered by the nature of AI, which most people forget is based on predicting human-like responses to prompts based on large amounts of past data. ChatGPT, for example, summarizes text by extracting its features and generating phrases that predict a possible summary of the content. However, it does not have real knowledge or true understanding of the context and is therefore not reliable. Likewise, exposing our schools – and workplaces – to a backward-looking, context-agnostic “intelligence” system is fraught with risk. Furthermore, creativity and human expression can only go so far. Yes, AI is here to stay, but it has to be on our terms. In 2024, educators must find productive ways to define and integrate the use of AI in the classroom rather than banning it as forbidden fruit.
James Evans, Max Palevsky Professor of Sociology and Director of the Knowledge Lab
OpenAI's ChatGPT, built on Google's machine learning Transformer architecture, wowed the world in late November 2022, attracting more than 100 million unique users within two months. ChatGPT underlies GPT-4, a massive language model with 1.7 trillion parameters that took nearly 100 days and $100 million to train. Some competitors are close behind, including Claude from Anthropic (backed by Amazon) and Gemini from Deep Mind (100% owned by Google). The sudden success of these models has transformed the “Turing test” (Turing 1950), which assesses an algorithm's ability to successfully mimic human interaction, from an unreasonable expectation to a fundamental expectation. Since then, new AI services have emerged daily to automate and transform human tasks, ranging from computer programming to journalism to art, science and invention. With this technology, AI is transforming both routine and creative tasks and promises to transform the future of work. These models have enabled automated translations of government documents, but also seamlessly blend fact and fiction, and have become linchpins in persuasion campaigns ranging from commercial advertising to political misinformation and disinformation initiatives. Increasingly, these models are used to control, execute, and synthesize the output of other tools, including databases, programming languages, and first-principles simulators.
The size and scope of these so-called “foundation models,” which can be tailored to specific tasks, prohibit all but the largest major technology companies from building from the ground up. This is likely to lead to an oligopoly in the AI services market in the medium term, resulting in the stable dominance of a few dominant models (e.g. “Coke or Pepsi?”) that attract the lion's share of attention, with countless small, special-purpose ones AI models at the edge. In the coming year, we will see a number of attempts to capitalize on these models, identify their “killer apps” and integrate them into established commercial pipelines such as online advertising. As these and related models are increasingly used to make important societal decisions that impact human well-being (e.g., judicial decisions, hiring, college admissions, research investments, grant disbursements, capital allocations), governments are increasing their exploration intensify AI regulation.
I increasingly expect governments to build checks and balances into a diverse ecology of AI algorithms. You will soon realize that we can only control powerful commercial AIs if other AIs are adversarially designed to examine, regulate, discipline and govern them. Finally, we will see new services such as intelligent search and native language-driven automation quickly emerge for society. Powerful language agents also become significantly more persuasive, blurring the line between credible facts and motivated opinions. Thinning the resulting information thicket will likely require a coordinated effort between (1) corporate initiatives to certify the information produced by their AIs (e.g., as early search engines effectively did for pornography), (2) government policy requirements, and (3) new ones Markets for personalized information agents and adaptive filters that enable tailored and potentially protected and polarized information environments.
Aziz Huq, Frank and Bernice J. Greenberg Professor of Law
Since the invention of the backpropagation algorithm in the 1990s, modern AI and the data circulation on which it is based have been only loosely regulated: the result is a flourishing of start-ups, supposed killer apps and unicorns. Of course, the state was never completely absent – as a financier, customer and silent partner. But as a regulator, it was most notable for its silence.
This impression has been somewhat misleading for some time, and 2024 will show that regulation is already here and is becoming ever more extensive and consequential. This is true even if the most hotly debated efforts to hold AI developers accountable for spillover harms such as bias and invasions of privacy have not yet had much practical impact. Because these efforts are perhaps also the most consequential. Because if the regulatory landscape is really changing, it will be because of more subtle, tectonic shifts.
On the surface, the regulatory landscape is becoming increasingly divided depending on jurisdiction. In December 2023, the Parliament and the Council of the European Union reached a preliminary agreement on a comprehensive AI law. In 2021, 2022 and 2023, the Chinese government issued separate tranches of AI regulation, targeting individual topics such as generative AI, taking a comprehensive approach with quiet but relentless confidence. In the United States, Biden's executive order on AI largely regulates the government directly, but could begin to impact the private sector due to the federal government's extensive procurement activities. At the same time, polarization in Congress makes comprehensive regulation highly unlikely.
But beneath the surface, the United States remains a highly influential global regulator that simply uses more subterranean tools. The focus is on export controls for semiconductors and the equipment required to produce them for 2022 and 2023, most of which are aimed at China. Of course the latter reacted accordingly. As this trade war intensifies – and potentially peaks under a new Trump administration – markets for the raw materials that make up AI will come under increasing pressure. And unlike direct regulation, there are no legal challenges or workarounds to soften the blow.
Indirectly, if not directly, the world of entrepreneurs who do whatever they want is coming to an end.
Articles reflect the opinions of their authors, not necessarily those of the University of Chicago, the Booth School of Business, or their faculty.
Comments are closed.