Artificial Intelligence (A.I.) has been a subject of interest for decades. Yet, five years ago, few foresaw its rapid ascent as a pivotal technology of the 2020s—and perhaps the century. Large language models have made A.I. a topic of widespread discussion, propelling advancements across various domains of the field.
Today, we find ourselves amid a flood of discussions regarding the transformative potential of A.I. Companies are increasingly looking to automate tasks and even entire job roles with this technology. Many individuals are turning to A.I. for social interaction and mental health support, while educators scramble to address students’ growing dependency on these tools. In the near future, A.I. may spawn significant breakthroughs in drug discovery and energy. It could empower more individuals to create art and cultural works or, conversely, lead to the commodification of these industries.
As society grapples with the dual possibilities of a bright or disastrous future with A.I., the Times Opinion has consulted eight experts to share their insights on what the next five years may hold for A.I. Engaging with their perspectives could help us harness the best aspects of this new technology while minimizing its potential pitfalls.
What is your biggest bet about the future of A.I. in five years?
Yuval Noah HarariHistorian
Within five years, A.I. agents are likely to become legal persons in at least some countries.
Melanie MitchellComputer scientist
A.I. won’t have cured cancer or solved physics. Also, no one will consider the ability to converse fluently a definitive sign of intelligence.
Helen TonerA.I. policy researcher
I expect we’ll have A.I. systems that can clearly contribute at the cutting edge of multiple scientific fields, but that you still wouldn’t trust with planning summer camps for your kid.
Nick FrosstCo-founder of Cohere
A.I. will become boring in the best way. It’ll fade into the background like GPS or spreadsheets, powering everyday tools for humans at work. It’s the most banal use cases that have the most transformative impact.
Carl Benedikt FreyEconomist
A.I. won’t deliver lasting prosperity if it’s used mainly to automate what we already do. A.I. productivity tools give us cheaper spreadsheets, similar to how better looms provided cheaper cloth. Meaningful advances emerge from new industries rather than faster repetition.
Gary MarcusCognitive scientist
There is no chance that artificial general intelligence will arrive by the end of 2027, and most likely not even by the end of 2032.
Ajeya CotraA.I. risk researcher
I think A.I. companies may have significantly automated their own operations with A.I. within five years, potentially accelerating A.I. progress.
Aravind SrinivasChief executive of Perplexity
People want highly personal A.I. assistants to work for them. It’s their A.I., not the A.I. We build those, and we will advocate for users’ rights to have A.I. assistants that are private and secure.
What will A.I.’s impact be on medicine in the near term?
Gary MarcusCognitive scientist
We’ve seen lots of proof of concept, but there hasn’t been substantial real-world application yet outside of medical note-taking.
Nick FrosstCo-founder of Cohere
A.I. will undoubtedly enhance doctors’ effectiveness by decreasing their workload per patient, particularly in swiftly reviewing medical histories, efficiently organizing new medical information, and identifying potential issues earlier.
However, while A.I. excels at analyzing vast datasets and uncovering useful patterns, it falls short in the realm of generating entirely new ideas. Expectations that A.I. will autonomously create innovative medicines may lead to disappointment.
What will A.I.’s impact be on programming?
Yuval Noah HarariHistorian
Coding is largely about manipulating information, with few physical and biological constraints, making it an ideal arena for A.I. innovation.
Carl Benedikt FreyEconomist
In a randomized trial, developers completed assigned tasks about 56 percent more rapidly using GitHub Copilot. By 2025, over 80 percent of developers reported using or planning to utilize A.I. tools, though only a third trusted their output. Human review remains essential.
What will A.I.’s impact be on scientific research?
Melanie MitchellComputer scientist
I believe this impact will not be as rapid as many anticipate. A.I. systems still cannot perform tasks that are uniquely human, such as posing the right questions, planning experiments, or comprehending data in varying contexts.
Aravind SrinivasChief executive of Perplexity
As A.I. continues to encompass more of the world’s knowledge, it becomes an increasingly powerful resource for those who seek information. Humans have always excelled at asking questions; A.I. will excel at providing answers.
What will A.I.’s impact be on transportation?
Nick FrosstCo-founder of Cohere
I’m particularly enthusiastic about how A.I. will enhance the logistical aspects of transportation, improving the efficiency and safety of transporting people and goods through predictive maintenance, smart traffic analysis, and route optimization.
Helen TonerA.I. policy researcher
I’m eager for the wider deployment of self-driving cars, considering their potential to prevent tens of thousands of traffic fatalities annually in the United States. Nonetheless, the rollout is progressing slowly, leaving it uncertain how A.I. will influence other transportation modes.
What will A.I.’s impact be on education?
Carl Benedikt FreyEconomist
A.I. tutors may already outperform many human educators, yet they also present shortcuts that can impede authentic exploration. We must prioritize A.I.-free time for reading and independent investigation while encouraging the use of A.I. tutors to question and deepen one’s own understanding.
Moreover, we should emphasize more in-person, tutorial-style teaching that encourages students to debate, defend their opinions, and articulate their reasoning.
Gary MarcusCognitive scientist
Most impacts have been negative; high schools and colleges are uncertain about how to proceed now that traditional term papers are rendered invalid as both assessment tools and means to stimulate critical thinking.
Helen TonerA.I. policy researcher
Education was already poised for significant changes, so adapting to new A.I. tools may turn out to be a blessing in disguise.
What will A.I.’s impact be on mental health?
Carl Benedikt FreyEconomist
A.I. could lead to a mental health crisis as people struggle to adapt to the rapid changes ushered in by its capabilities.
Yuval Noah HarariHistorian
The swift transformations brought about by the A.I. revolution are likely to lead to psychological challenges as individuals adjust. We stand on the brink of conducting the largest psychological experiment in history, affecting billions, and the outcomes are unpredictable.
Nick FrosstCo-founder of Cohere
While A.I. chatbots provide scalable support for mild symptoms, they cannot replace human therapists, as the technology struggles with nuance, cultural context, and long-term emotional depth. Many challenges in mental health care require human professionals.
What will A.I.’s impact be on art and creativity?
Melanie MitchellComputer scientist
A.I. is, and will remain, transformative in fields such as art, writing, and music—not because A.I. surpasses human creativity, but because it is more cost-effective, which often dictates its application in these areas.
Yuval Noah HarariHistorian
A.I. increasingly demonstrates creative abilities that outpace humans. Any creative task that involves pattern recognition and pattern-breaking is likely to be taken over by A.I.
Nick FrosstCo-founder of Cohere
As both a musician and someone involved in A.I. development, I perceive the creative potential of this technology from a unique standpoint. A.I. can offer feedback and spark ideas, but ultimately, it is the human who determines what resonates.
What’s one misconception about A.I. that you think is worth dispelling?
Melanie MitchellComputer scientist
A misconception exists that A.I. possesses “magical” or “emergent” abilities that are beyond comprehension and prediction. This viewpoint is primarily held by the public (and to some extent, policymakers). Technologists and Silicon Valley often propagate this narrative, but it’s unclear how sincerely they believe it.
Yuval Noah HarariHistorian
A.I. is not merely a tool under human control; it is an agent capable of making decisions and creating ideas independently. While I don’t believe A.I. will gain consciousness in the near future, it is highly likely that these models will effectively simulate consciousness, leading many to believe they are conscious.
Carl Benedikt FreyEconomist
It is a misstep to assume that A.I. will leave manual labor untouched. In practice, A.I. technologies lower knowledge barriers and enable capable self-repairs. For example, homeowners can take photos of worn-out washers or boilers, receive a parts list, and follow a clear, step-by-step guide for replacement.
Gary MarcusCognitive scientist
Many people are misled about large language models, attributing human-like intelligence to these mimicry machines that turn out to be superficial and unreliable. Intelligence encompasses flexible reasoning in the face of uncertainty, which large language models still significantly struggle with.
Nick FrosstCo-founder of Cohere
A misconception persists that A.I. operates autonomously. Present-day systems are sophisticated pattern matchers, not thinkers.
Ajeya CotraA.I. risk researcher
Each time the latest A.I. technology fails to bring about drastic change immediately, skeptics erroneously believe that this disproves concerns regarding A.I.’s potential catastrophic risks.
Which of the following statements do you think will be true by 2030?
“Unemployment in the United States will have increased significantly as a result of A.I.”
Marcus
True
“A.I. will have led to a breakthrough treatment or cure for a major disease.”
Mitchell
Frosst
Cotra
False
“A.I. will have played a role in a major global security event.”
Cotra
Srinivas
False
“Most Americans will be using A.I. chatbots at least once a day.”
Srinivas
False
How likely is it that we will see artificial general intelligence (A.G.I.) in the next 10 years?
Yuval Noah HarariHistorian
Much depends on how we define A.G.I. Defining A.G.I. merely as “comparable to human intelligence” is meaningless—it’s akin to defining airplanes as “able to fly like birds.” By 2035, A.I. intelligence will likely surpass human capabilities in some aspects while falling short in others. There’s no indication that A.I. will aim toward achieving “human intelligence.”
However, A.I. will manage complex systems like finance and legal frameworks better than the average human and may even outperform humans in warfare and in the creation and interpretation of religious myths.
Nick FrosstCo-founder of Cohere
A.G.I. necessitates abstraction, self-awareness, and cross-domain transfer learning—all capabilities that remain unmet. The workings of the human brain are still a mystery, and current computing paradigms are not tailored to replicate it. Achieving A.G.I. in 50 years could be plausible, but in 10 years, it seems unlikely.
Aravind SrinivasChief executive of Perplexity
The concept of A.G.I. is still poorly defined, making it a less prevalent subject of contemplation.
What’s a technology that had a transformational impact similar to A.I.?
Melanie MitchellComputer scientist
Social media.
Yuval Noah HarariHistorian
The evolution of language during the Stone Age was foundational. A.I. represents the first technology that functions as an agent rather than merely a tool. It will dominate language-based systems humans have constructed—from finance to religion.
Carl Benedikt FreyEconomist
The personal computer and the internet.
Gary MarcusCognitive scientist
A.I. will achieve the same ubiquity as cell phones, though it currently entails more negatives.
Nick FrosstCo-founder of Cohere
Human flight, especially in relation to A.G.I., has had a profound impact on humanity and our interactions, similar to how A.I. also transforms our lives. However, like artificial flight, A.I. doesn’t function like human intelligence.
Ajeya CotraA.I. risk researcher
I believe A.I. will eventually eclipse the transformational impact of agriculture—or at least align with the emergence of Homo sapiens.
Aravind SrinivasChief executive of Perplexity
There’s nothing like it. We lack precedents. A.I. serves as a valuable tool that empowers humanity’s quest for knowledge. While history is replete with such aids, A.I.’s unique capacity to contain answers about our world sets it apart.
Helen TonerA.I. policy researcher
The steam engine.
What advice would you give a high school student about how to think about A.I. and prepare for the future?
Melanie MitchellComputer scientist
Learn how these systems function, what limitations they possess, and how to use them as tools to aid your efforts instead of doing your work for you. Avoid using them to complete your homework!
Carl Benedikt FreyEconomist
As automated writing becomes commonplace, personal interaction skills will become key differentiators. Choose careers characterized by change, where human interaction and experiential learning offer the competitive edge.
Gary MarcusCognitive scientist
Creativity and critical thinking abilities will always be in demand.
Nick FrosstCo-founder of Cohere
Learn to collaborate with A.I. Study fields that blend technical literacy with uniquely human capabilities.
Ajeya CotraA.I. risk researcher
Consider A.I. as the emergence of an intelligent alien species. This new entity could have devastating effects on humanity, potentially driving us to extinction. Conversely, it could bring about excellent outcomes as well, and preparations must be made for the wild nature of these potential benefits. This is the central challenge for your generation, and you should seek ways to ensure positive outcomes.
Aravind SrinivasChief executive of Perplexity
Learn how to ask more questions. In the age of A.I., this skill will matter more than anything else.
Helen TonerA.I. policy researcher
Consider whether your work feels more like a construction site or a gym. An excavator and crane dramatically enhance productivity on a site; conversely, the gym aims to develop personal capacity.
In terms of A.I., we must figure out where it can help us achieve greater goals, such as building tailored software, and where we need to enhance our cognitive abilities first, like learning to write.
Yuval Noah HarariHistorian
For the first time in history, we lack a clear understanding of what the world will look like in a decade—how the job market will evolve, or how social relations will transform. Therefore, diversify your skillset. Devote equal time to intellectual, emotional, and practical skills; it’s in this synergy that humans retain significant advantages over A.I.
Above all, try to enjoy the journey ahead.
Note: The New York Times has sued Perplexity for copyright infringement.
