Friday morning at SXSW, Niall Firth, the Executive Editor of the Newsroom at the Massachusetts Institute of Technology (MIT), reviewed MIT’s annual list of 10 breakthrough technologies for 2025. The review is particularly interesting since it comes from MIT, an independent institution with a 135-year history and a sterling reputation. MIT researchers analyze reports, speak with experts, and closely track technological advancements to predict which innovations will have the most significant impact on our future.
MIT identifies emerging technologies based on two key factors: their long-term significance and their potential to become commercial realities. According to Firth, compiling this list involves extensive discussion, debate, and, at times, outright argument.
The Top Ten Breakthrough Technologies for 2025
Here are the ten technologies that made this year’s list:
- Vera C. Rubin Observatory – This observatory, based in Chile, will conduct a ten-year survey of the night sky, creating the most detailed digital map ever. It aims to uncover new galaxies, explore black holes, and study dark matter.
- Generative AI Search – Perhaps it is stating the obvious, but Gen AI is a major shift in how we navigate the internet. Unlike traditional search engines, generative AI responds in a conversational style, offering more direct answers. This could be a step toward AI agents capable of completing complex tasks (BUT… see below).
- Small Language Models – Unlike large AI models trained on massive datasets, small language models require significantly less data and computing power, making them more efficient and adaptable for specific tasks.
- Reducing Cattle Burps – Cattle burps are a significant source of methane emissions. Scientists are developing supplements to reduce methane output, helping address one of agriculture’s largest environmental challenges.
- Robotaxis – Robotaxis are exploding with more and more use in more and more cities.Autonomous taxis are collecting valuable data, improving with each deployment. Companies like Waymo are leading the charge, with AI enabling cross-city learning and adaptation.
- Cleaner Jet Fuel – Technological advancements have led to new ways of producing jet fuel with significantly lower emissions, marking a major step in sustainable aviation.
- Fast-Learning Robots – Previously, robots could only be trained for a single task. Now, thanks to new training techniques and large language models, robots can watch and learn from humans, opening the door to more sophisticated automation.
- Long-Acting HIV Prevention Medications – A breakthrough in medical treatment that offers long-term HIV prevention, reducing the need for daily medication.
- Green Steel Technologies – Innovations in steel production that reduce carbon emissions, making manufacturing more sustainable.
- Lab-Grown Stem Therapies – Scientists have struggled to create useful stem cell therapies in the past. New advancements may finally make lab-grown stem treatments viable for conditions like diabetes and epilepsy.
Technologies That Almost Made the Cut
Firth also highlighted a few technologies that didn’t make the final list but were close contenders:
- Virtual power plants
- AI agents
- EV flying taxis – While self-driving taxis are one thing, fully autonomous airborne taxis still seem like a leap too far. (I’m not quite ready to get in a Waymo but get in a pilot-less airborne taxis? Don’t think so).
Why This Matters for Law and Legal Tech
Firth emphasized that one of the key criteria for selecting technologies is their broader relevance — what problem do they solve? Here’s how some of these breakthroughs could impact the legal industry:
- Small Language Models and Legal AI – Unlike large AI models trained on vast public datasets, small language models can be built on private, secure datasets, making them ideal for legal applications. Law firms and in-house legal teams could develop AI tools trained on their own cases and internal documents, improving efficiency while maintaining confidentiality. These models also require far less computational power, making them more practical and cost-effective.
Firth pointed out that the gains from large models are plateauing, while small models are becoming more powerful and adaptable. This shift is already underway, with some publicly available models being scaled down to work on specific, localized data.
Use of these models have lots of applications for law. They could be used on large e-discovery data sets. They could be used to access a law firm’s past efforts. They could mine clients data to provide answers to legal questions efficiently. For that matter, they could allow in house legal to answer questions from company data without engaging outside counsel on certain issues.
As I have written before, legal Gen AI providers like vLex are trying to marry the late public data sets with internal documents of large firms and businesses. But small language models that do even less but are responsive to limited needs could become more the norm.
- A Problem-First Approach to Legal Tech – One of the key takeaways from the MIT list is that breakthrough technology isn’t just about innovation—it’s about application. Instead of chasing the latest tech trends, legal professionals should start by identifying real pain points in the practice of law and then determine how technology can address them. This is how we get advancements like green steel, medical breakthroughs, and methane-reducing cattle feed. If we can solve those problems, we can use technology to solve inefficiencies in legal workflows. (Yes, if they come up with ways to eliminate the BS, that can certainly have applications to legal!)
- AI Agents Are Still a Work in Progress – AI agents — systems designed to act autonomously like human assistants — didn’t make MIT’s list, and for good reason. While there’s been a lot of hype around AI agents automating complex workflows, Firth made it clear that the technology isn’t ready yet. OpenAI has been trying to create AI agents, but so far, they haven’t been able to make them reliable. One major challenge is trust: AI agents need access to extensive user data to function effectively, which raises significant privacy concerns. In legal, where confidentiality is paramount, this remains a major roadblock.
A Final Thought
MIT’s list, like much of what’s being discussed at SXSW, highlights the gap between technological advancements and the practical realities of implementation. Some innovations, particularly in sustainability and healthcare, may now face regulatory or commercial hurdles before they reach widespread adoption. While that’s a challenge, it’s also encouraging that SXSW continues to foster conversations about science, technology, and their potential applications.
For the legal industry, the takeaway is clear: The future isn’t just about adopting new technology — it’s about strategically applying it to solve the right problems. And as small language models, AI-driven automation, and other advancements evolve, firms that adapt thoughtfully will be the ones best positioned for success.
Stephen Embry is a lawyer, speaker, blogger and writer. He publishes TechLaw Crossroads, a blog devoted to the examination of the tension between technology, the law, and the practice of law.
The post MIT Reveals 2025 Breakthrough Tech At SXSW: What It Means For Legal appeared first on Above the Law.
Friday morning at SXSW, Niall Firth, the Executive Editor of the Newsroom at the Massachusetts Institute of Technology (MIT), reviewed MIT’s annual list of 10 breakthrough technologies for 2025. The review is particularly interesting since it comes from MIT, an independent institution with a 135-year history and a sterling reputation. MIT researchers analyze reports, speak with experts, and closely track technological advancements to predict which innovations will have the most significant impact on our future.
MIT identifies emerging technologies based on two key factors: their long-term significance and their potential to become commercial realities. According to Firth, compiling this list involves extensive discussion, debate, and, at times, outright argument.
The Top Ten Breakthrough Technologies for 2025
Here are the ten technologies that made this year’s list:
- Vera C. Rubin Observatory – This observatory, based in Chile, will conduct a ten-year survey of the night sky, creating the most detailed digital map ever. It aims to uncover new galaxies, explore black holes, and study dark matter.
- Generative AI Search – Perhaps it is stating the obvious, but Gen AI is a major shift in how we navigate the internet. Unlike traditional search engines, generative AI responds in a conversational style, offering more direct answers. This could be a step toward AI agents capable of completing complex tasks (BUT… see below).
- Small Language Models – Unlike large AI models trained on massive datasets, small language models require significantly less data and computing power, making them more efficient and adaptable for specific tasks.
- Reducing Cattle Burps – Cattle burps are a significant source of methane emissions. Scientists are developing supplements to reduce methane output, helping address one of agriculture’s largest environmental challenges.
- Robotaxis – Robotaxis are exploding with more and more use in more and more cities.Autonomous taxis are collecting valuable data, improving with each deployment. Companies like Waymo are leading the charge, with AI enabling cross-city learning and adaptation.
- Cleaner Jet Fuel – Technological advancements have led to new ways of producing jet fuel with significantly lower emissions, marking a major step in sustainable aviation.
- Fast-Learning Robots – Previously, robots could only be trained for a single task. Now, thanks to new training techniques and large language models, robots can watch and learn from humans, opening the door to more sophisticated automation.
- Long-Acting HIV Prevention Medications – A breakthrough in medical treatment that offers long-term HIV prevention, reducing the need for daily medication.
- Green Steel Technologies – Innovations in steel production that reduce carbon emissions, making manufacturing more sustainable.
- Lab-Grown Stem Therapies – Scientists have struggled to create useful stem cell therapies in the past. New advancements may finally make lab-grown stem treatments viable for conditions like diabetes and epilepsy.
Technologies That Almost Made the Cut
Firth also highlighted a few technologies that didn’t make the final list but were close contenders:
- Virtual power plants
- AI agents
- EV flying taxis – While self-driving taxis are one thing, fully autonomous airborne taxis still seem like a leap too far. (I’m not quite ready to get in a Waymo but get in a pilot-less airborne taxis? Don’t think so).
Why This Matters for Law and Legal Tech
Firth emphasized that one of the key criteria for selecting technologies is their broader relevance — what problem do they solve? Here’s how some of these breakthroughs could impact the legal industry:
- Small Language Models and Legal AI – Unlike large AI models trained on vast public datasets, small language models can be built on private, secure datasets, making them ideal for legal applications. Law firms and in-house legal teams could develop AI tools trained on their own cases and internal documents, improving efficiency while maintaining confidentiality. These models also require far less computational power, making them more practical and cost-effective.
Firth pointed out that the gains from large models are plateauing, while small models are becoming more powerful and adaptable. This shift is already underway, with some publicly available models being scaled down to work on specific, localized data.
Use of these models have lots of applications for law. They could be used on large e-discovery data sets. They could be used to access a law firm’s past efforts. They could mine clients data to provide answers to legal questions efficiently. For that matter, they could allow in house legal to answer questions from company data without engaging outside counsel on certain issues.
As I have written before, legal Gen AI providers like vLex are trying to marry the late public data sets with internal documents of large firms and businesses. But small language models that do even less but are responsive to limited needs could become more the norm.
- A Problem-First Approach to Legal Tech – One of the key takeaways from the MIT list is that breakthrough technology isn’t just about innovation—it’s about application. Instead of chasing the latest tech trends, legal professionals should start by identifying real pain points in the practice of law and then determine how technology can address them. This is how we get advancements like green steel, medical breakthroughs, and methane-reducing cattle feed. If we can solve those problems, we can use technology to solve inefficiencies in legal workflows. (Yes, if they come up with ways to eliminate the BS, that can certainly have applications to legal!)
- AI Agents Are Still a Work in Progress – AI agents — systems designed to act autonomously like human assistants — didn’t make MIT’s list, and for good reason. While there’s been a lot of hype around AI agents automating complex workflows, Firth made it clear that the technology isn’t ready yet. OpenAI has been trying to create AI agents, but so far, they haven’t been able to make them reliable. One major challenge is trust: AI agents need access to extensive user data to function effectively, which raises significant privacy concerns. In legal, where confidentiality is paramount, this remains a major roadblock.
A Final Thought
MIT’s list, like much of what’s being discussed at SXSW, highlights the gap between technological advancements and the practical realities of implementation. Some innovations, particularly in sustainability and healthcare, may now face regulatory or commercial hurdles before they reach widespread adoption. While that’s a challenge, it’s also encouraging that SXSW continues to foster conversations about science, technology, and their potential applications.
For the legal industry, the takeaway is clear: The future isn’t just about adopting new technology — it’s about strategically applying it to solve the right problems. And as small language models, AI-driven automation, and other advancements evolve, firms that adapt thoughtfully will be the ones best positioned for success.
Stephen Embry is a lawyer, speaker, blogger and writer. He publishes TechLaw Crossroads, a blog devoted to the examination of the tension between technology, the law, and the practice of law.
The post MIT Reveals 2025 Breakthrough Tech At SXSW: What It Means For Legal appeared first on Above the Law.
