Who will benefit from artificial intelligence? This fundamental question, which has become especially salient during the rise of AI over the past few years, took center stage at a conference at MIT on Wednesday as speakers and audience members grappled with various aspects of AI’s impact.
In one of the conference’s keynotes, journalist Karen Hao ’15 called for a change in the trajectory of AI development, including a shift away from massive scale-ups of data usage, data centers, and models used to develop tools under the name “artificial general intelligence.”
“This scale is unnecessary,” said Hao, who has become a prominent voice in the AI debate. “We don’t need AI and computing at this scale to realize the benefits.” In fact, she added, “If we truly want AI to have widespread benefits, we need to shift away from this approach quickly.”
Mr. Hao is a former staff member. wall street journal and MIT Technology Reviewauthor of the 2025 book Empire of AI. She has reported extensively on the growth of the AI industry.
In his talk, Hao outlined the astonishing scale of datasets that the largest AI companies are currently using to develop large-scale language models. He also highlighted some of the trade-offs in this scale-up, such as hyperscale data centers’ high energy consumption and emissions, as well as consuming large amounts of water. Based on his report, Hao also cited the human toll caused by gig economy workers around the world manually inputting data into hyperscale models.
In contrast, Hao suggested that another path for AI may exist, using the example of AlphaFold, a Nobel Prize-winning tool used to identify the structure of proteins. This represents the concept of “small, task-specific AI models that leverage AI’s computational power to tackle a wide range of problems,” Hao said.
“It is trained on a highly curated dataset that is only relevant to the problem at hand: protein folding and amino acid sequence. … Because the dataset is small and the model is small, fast supercomputing is not needed, but it still provides significant benefits.”
In the second keynote, scholar Paola Ricourt outlined a number of conceptual keys for assessing the usefulness of AI and emphasized the desirability of a purpose-driven approach to AI.
“Technology is meaningless if it’s not responsive to the community that will be using it,” Ricourt said.
She is a professor at Tecnológico de Monterrey in Mexico and a faculty assistant at the Berkman Klein Center for Internet and Society at Harvard University. Ms. Ricourt has also served on expert committees such as the Global Partnership for AI, UNESCO’s AI Ethics Experts Without Borders, and the Women in Ethical AI Project.
The event was sponsored by MIT’s Women’s and Gender Studies Program. Manduhai Bhuyandergar, director of the program and professor of anthropology, gave introductory remarks.
The event, titled “Gender, Empire, and AI: Symposium and Design Workshop,” was held in the conference space of the MIT Schwarzman College of Computing, and more than 300 people attended the keynote address. The event also had a section dedicated to discussion groups, with an afternoon session featuring design sessions in six different subject areas.
In his speech, Hao decried the vague nature of the AI debate, suggesting it was holding back a more thoughtful discussion about the direction of the industry.
“One of the challenges when talking about AI is that there is a complete lack of specificity in the term ‘artificial intelligence,'” Hao said. “It’s like the word ‘transportation.’ It can refer to anything from bicycles to rockets.” As a result, she said, “When you talk about leveraging that advantage, you actually have to be very specific. What AI technologies are you talking about? Which AI technologies do you want to leverage more of?”
In her view, smaller tools, similar to bicycles, are more useful on a daily basis. As another example, Hao mentioned the project Climate Change AI, which focuses on tools to help improve building energy efficiency, track emissions, optimize supply chains, predict extreme weather, and more.
“This is the vision of AI that we should strive to build toward,” Hao said.
Finally, Mr. Hao encouraged the audience to actively participate in AI-related discussions and projects, stating that the technology’s trajectory is still uncertain and public intervention is important.
Quoting writer Rebecca Solnit, Hao suggested to the audience, “Hope lies in the premise that we don’t know what will happen, and that hope lies in the premise that there is room for action in the breadth of uncertainty.” She also said, “Each of you has an active role in shaping technology development.”
Ricotte similarly encouraged attendees to actively participate in AI issues, noting that technology is most effective when it addresses the pressing daily needs of all citizens.
“We have a responsibility to make hope possible,” Ricourt said.
