Exclusive Interview: How AI will disrupt the legal industry
June 14, 2024
Good morning, and happy Friday. I spoke with the CEO of legal AI firm Robin AI about reliable AI solutions and where he thinks the industry is headed.
Read on for the full story.
In today’s newsletter:
🌲 AI for Good: Another tool to prevent wildfires
🔬 OpenAI Briefing: CTO says the models in the lab aren’t that advanced
⛔️ Google researcher says OpenAI & ChatGPT are an ‘off-ramp to AGI’
🏛️ Exclusive Interview: How AI will disrupt the legal industry
AI for Good: Another tool to prevent wildfires
Photo by Sebastian Unrau (Unsplash).
California’s 2020 wildfire season was the largest wildfire season in the modern history of the state. Nearly 10,000 fires burned across 4.3 million acres of land.
A team of researchers recently applied deep learning methods to study the makeup of California’s forests before that season, and they found a culprit: Tons of dead trees lying among the living.
The details: The researchers applied deep-learning dead tree detection techniques to a series of sub-aerial photographs from 2020.
They found 91.4 million dead trees over 27.8 million hectares of vegetated area. 60% of those dead trees were in groups of living trees.
The researchers added that mapping and properly estimating levels of dead trees is a “critical input” for monitoring fire risk.
Why it matters: The researchers said that further mapping could “contribute to early warning systems in areas that may face widespread forest mortality in a changing climate.”
OpenAI Briefing: CTO says the models in the lab aren’t that advanced
Image Source: OpenAI
OpenAI CTO Mira Murati told Fortune Wednesday that the company isn’t hiding AGI somewhere in its lab.
“Inside the labs, we have these capable models. And they’re not that far ahead from what the public has access to for free,” she said, adding that this “allows people to prepare for the advent of bringing advanced AI into the world.”
OpenAI last month began training its next frontier model, saying that it anticipates “the resulting systems to bring us to the next level of capabilities on our path to AGI.”
“I don’t think people quite realize the significance of this moment, so I will spell it out: OpenAI has been bluffing all along, and their CTO just gave away the fact that they are not holding aces,” cognitive scientist Gary Marcus said.
In other OpenAI News: The Information reported that neither Apple nor OpenAI are paying each other for the coming integration of ChatGPT into Apple’s OS.
OpenAI’s hope is to drive people to the paid version of ChatGPT. Apple will take its customary 30% slice of any of those purchases.
This is a markedly different arrangement than Apple’s search engine partnership with Google, in which Google paid Apple billions of dollars to be Safari’s default option. This search engine dominance later landed Google in hot water with the U.S. Justice Department for alleged monopolistic behavior.
Together with DataChat
The conversation is the code
Creating data analytics that can transform your company’s strategy shouldn’t require coding expertise.
With DataChat, it doesn’t. That’s why we at The Deep View, along with other SMBs and Fortune 100s, use DataChat. With DataChat, you don’t have to be an expert Data Scientist or know anything about Python, SQL, or coding to get insights from your data. DataChat uses built-in data science and GenAI to answer your questions.
Business users: Just ask questions in plain English and instantly receive actionable analytics and insights on your data.
Analysts: Work in a familiar spreadsheet UI that includes no-code machine learning and predictive analytics.
DataChat never shares your data with LLMs. It also prioritizes transparency, documenting each step it takes ensuring your team can validate results and understand its thought process.
Unlock the value of your data today with DataChat. Get started today.
Researcher says OpenAI & ChatGPT are an ‘off-ramp to AGI’
Photo by David Martin (Unsplash).
Francois Chollet, a deep learning researcher at Google, said earlier this week that OpenAI “basically set back progress to artificial general intelligence (AGI) by 5-10 years.”
“They caused this complete closing down of frontier research publishing,” he said. “Now LLMs have sucked the oxygen out of the room … I see LLMs as more of an off-ramp on the path to AGI.”
Chollet’s view, and the view of a growing number of researchers (Gary Marcus, Yann LeCun, Grady Booch, etc.) is that any “reasoning” LLMs might demonstrate is nothing more than an illusion, as it is a direct result of their training data. Chollet’s goal is to build systems that demonstrate real reasoning capabilities — in other words, systems that can solve novel problems not included in their training data.
“Scale is not intelligence,” Chollet said. “General intelligence is the ability to approach any problem, any skill, and very quickly master it, using very little data.”
He added that LLMs are nothing more than information retrieval systems, whereas human intelligence is so much more: “We have a memory, but we are not a memory.”
Chollet at the same time launched Arc Prize, a $1 million competition to try and get AI research “back on track towards AGI.”
💰AI Jobs Board:
Associate Director, AI & ML Solutions : Johnson & Johnson · United States · New Jersey · Full-time · (Apply here)
Senior Manager Generative AI Security: AWS · United States · Multiple Locations · Full-time · (Apply here)
Senior Director, Data Science & AI: Pfizer · United States · Multiple Locations · Full-time · (Apply here)
🕋 Events: *
Can’t Miss: Join Nvidia, OpenAI and Google at AIQCon
Time is running out for you to snag tickets to attend Kolena’s AI Quality Conference (happening in San Francisco on June 25). In this jam-packed event, you’ll get the chance to engage with experts from the world’s biggest names in AI.
Learn, network and have fun doing it. TDV readers can purchase tickets here for 50% off using TDV50.
🌎 The Broad View:
OpenAI’s annualized revenue reaches $3.4 billion (The Information).
The technology that could force U.S. drivers to stop speeding (The Verge).
OpenAI significantly expands team of international lobbyists (Financial Times).
Microsoft President Brad Smith grilled by Congress after security breaches (Washington Post).
*Indicates a sponsored link
Together with Brave
Feed Your AI with Real Data
The Brave Search API brings affordable developer access to the Brave Search Index: an independent index of the Web with over 20 billion webpages.
What makes it special?
It’s built on real page visits from anonymous humans, which filters out tons of junk data.
It’s independent and proprietary, meaning no Big Tech biases or extortionate prices.
It’s refreshed daily, so it’s always accurate and up-to-date.
The Brave Search API can be used to assemble a dataset to train your AI models, and help with retrieval augmentation at the time of inference. And it does so affordably.
Integrating the Brave Search API into your workflow translates to more ethical data sourcing and more human-representative datasets.
Exclusive Interview: How AI will disrupt the legal industry
Photo by Sebastian Pichler (Unsplash).
Lawyers — costing an average of around $300 per hour — are notoriously expensive.
And for the vast majority of people, that cost is out of reach; Gillian Hadfield, a law professor and economist at the University of Toronto, estimated in 2020 that “well over 80%” of Americans who need legal help can’t get it because of the cost.
That environment seems tailor-made for disruption via generative AI. But there have already been hallucinatory mishaps, where models produced output that looks real but isn’t. In perhaps the most highly publicized example, two New York lawyers were sanctioned last year after they submitted a brief — written by ChatGPT — that featured fake cases.
Stanford studies have found that general-purpose models hallucinate on legal queries more than half the time, and even specially-designed legal models hallucinate around 20% of the time.
In the face of this challenge, the effort, according to Robin AI CEO Richard Robinson, is to be more thoughtful in how the tech is applied to the legal sector.
Robinson, himself a former lawyer, founded the company in 2019 specifically as a solution to the world of contract law. His intention was to create a human-in-the-loop system that would speed lawyers up (thereby reducing their cost) while also democratizing legal knowledge to the masses.
“The idea was you’re Batman, and we’re Robin,” Robinson told me.
He said that people don’t always need access to a high-powered lawyer. Sometimes, they just need a second set of eyes — a Lite lawyer, if you will — to sift through a new contract with an employer, or a landlord. “And these systems are really good at that.”
The guardrails: Robinson said that an important focus involves clearly clarifying the limits of such models: They won’t provide personalized legal advice and they won’t get people out of jail. They will scan contracts. With these limits established, and in the more narrow arena of contract review, Robinson isn’t too concerned about hallucinations.
“I still think it’s better than going on Google, or just not knowing,” he said. “I still think it’s better than having no help at all.”
‘First thing we do is, let’s kill all the lawyers’ (Shakespeare): Robinson expects that with AI, the bulk of legal work will eventually become enormously streamlined and accessible. He argued that lot of legal knowledge isn’t much more than “memorization,” something “humans have never actually been good at.”
“What will happen is something like 60% of the work will just stop needing people to be involved at all, because actually, it's fairly straightforward,” he said. “It's fairly predictable, it’s fairly consistent over time.”
But he believes that humans will always be involved in the parts of the legal process that aren’t straightforward: Making judgments, determining truth, assessing ambiguity, etc.
The result of this would likely be fewer, but better, lawyers.
“I don’t think the industry should be obsessed by how many lawyers there are,” Robinson said. “Its job is to serve society. I think the industry has had a really bad reputation for a long time, and I think these technologies give us a chance to chart a different course.”
Have cool resources or tools to share? Submit a tool or reach us by replying to this email (or DM us on Twitter).
*Indicates a sponsored link
SPONSOR THIS NEWSLETTER
The Deep View is currently one of the world’s fastest-growing newsletters, adding thousands of AI enthusiasts a week to our incredible family of over 200,000! Our readers work at top companies like Apple, Meta, OpenAI, Google, Microsoft and many more.
One last thing👇
That's a wrap for now! We hope you enjoyed today’s newsletter :)
What did you think of today's email? |
We appreciate your continued support! We'll catch you in the next edition 👋
-Ian Krietzberg, Editor-in-Chief, The Deep View