I admit\u00a0I\u2019ve been a skeptic.\u00a0But the press\u00a0conferences\u00a0and keynotes by the big players like\u00a0Nvidia<\/a>,\u00a0Samsung<\/a>,\u00a0LG<\/a>,\u00a0and a smaller\u00a0player,\u00a0AGI Inc.<\/a>,\u00a0suggest\u00a0that agentic is not just some pie in the sky\u00a0hyperbole\u00a0but a real thing that can do real things\u00a0for\u00a0real\u00a0people.<\/p>\n As\u00a0Amit Jain<\/a>, CEO of\u00a0Luma\u00a0AI<\/a>,\u00a0put it\u00a0as a guest in\u00a0the AMT\u00a0Keynote, \u201c2026 will be the year of the agents.\u201d<\/p>\n I\u2019m\u00a0now\u00a0convinced that agentic AI is enough of a thing for legal to begin thinking about its benefit and risks.<\/p>\n What\u00a0<\/strong>Agentic\u00a0<\/strong>AI May Mean<\/strong><\/p>\n What all\u00a0these\u00a0keynote\u00a0speakers talked about was the\u00a0potential for\u00a0agentic AI to create an honest to God assistant. Not a dumb GenAI assistant that tries to answer\u00a0just\u00a0what you ask, albeit often\u00a0unsuccessfully. But a real assistant who answers you, makes suggestions for next steps, helps you think\u00a0through\u00a0solutions, and helps you\u00a0implement them. It\u2019s like\u00a0having\u00a0a\u00a0devoted\u00a0friend\u00a0at your side whose only job is to help in every way.\u00a0<\/p>\n Some\u00a0consumer\u00a0examples\u00a0mentioned by\u00a0CES speakers\u00a0demonstrate what\u00a0agentic\u00a0AI\u00a0may\u00a0do\u00a0for\u00a0business.\u00a0An\u00a0agent can decipher what\u2019s in your\u00a0fridge\u00a0and\u00a0create a\u00a0recipe suggestion based on what\u2019s there and what you have historically liked. It can figure out how to season the food based on your taste. It can help you plan an anniversary night out with your spouse based on likes and dislikes it\u00a0knows.\u00a0<\/p>\n It can call you an Uber based on\u00a0your\u00a0request to get you an Uber. And then provide you with all the information you need\u00a0to catch the Uber. No more wandering around airport trying to figure out where and\u00a0how to catch your Uber.<\/p>\n It\u00a0doesn\u2019t take much imagination to think of\u00a0examples\u00a0that may aid\u00a0business and even legal.<\/p>\n I\u2019m sure you\u2019re thinking, as\u00a0I was, it\u2019s the\u00a0CES\u00a0same old, same old. Wild claims backed by extravagant productions with\u00a0heartwarming\u00a0videos about how AI\u00a0is\u00a0going to change our world. But this felt a little different once I reflected on a couple of\u00a0my own\u00a0recent\u00a0experiences\u00a0and\u00a0after\u00a0I heard Jensen Huang<\/a>, founder and CEO of\u00a0Nvidia,\u00a0speak.<\/p>\n Some\u00a0<\/strong>Personal Examples<\/strong><\/p>\n I got a glimpse of\u00a0what agentic AI could do\u00a0when I recently tried OpenAI\u2019s web browser,\u00a0ChatGPT\u00a0Atlas<\/a>. I asked\u00a0it\u00a0to help me pick flights to Las Vegas for CES. It accurately sorted through the options\u00a0on several websites, directed me to the\u00a0best one,\u00a0and with my permission, booked the flight\u00a0on that site. It saved me a bunch of\u00a0clicks,\u00a0and it\u00a0worked\u00a0pretty seamlessly. I\u2019m pretty sure with some more use, it would learn I only want direct flights and\u00a0would\u00a0like an aisle seat.\u00a0Admittedly, there have been\u00a0criticisms\u00a0of\u00a0Atlas, but\u00a0its capabilities I witnessed suggests its potential.<\/p>\n As for the notion that this is all pie in the sky and will never happen, the same has been said for\u00a0self-driving\u00a0cars. I\u00a0bought\u00a0one a few years ago and while the\u00a0self-driving\u00a0features at that time were okay, they weren\u2019t something I used every day or relied on.\u00a0But the\u00a0improvements\u00a0since then\u00a0are remarkable.\u00a0They are now so good,\u00a0I use\u00a0them\u00a0every day\u00a0on\u00a0just about every trip. And as\u00a0Jensen\u00a0pointed out\u00a0in his keynote, much of what is going on with these\u00a0vehicles\u00a0is a form of\u00a0and powered by\u00a0agentic AI.<\/p>\n The notion\u00a0that we\u00a0may be on the cusp of something\u00a0big\u00a0was\u00a0reiterated\u00a0when Jensen explained why\u00a0this may be the case.<\/p>\n Jensen\u2019s Keynote<\/strong><\/p>\n Jensen explained why agentic AI\u00a0may be\u00a0ready to\u00a0become mainstream.\u00a0Here\u2019s\u00a0why:<\/p>\n \u2022\u00a0Computing power has grown\u00a0exponentially.<\/p>\n \u2022\u00a0That growth has enabled AI programs to\u00a0understand and grasp data in things like PDFs, images, and audio files.<\/p>\n \u2022\u00a0AI programs are seeking\u00a0solutions\u00a0from multiple LLM and cloud servers, increasing the amount of data from which to cull answers.<\/p>\n \u2022\u00a0AI programs can now simulate\u00a0solutions for\u00a0situations for which there is an absence of data by\u00a0recognizing patterns from similar situations and understanding outcomes.<\/p>\n What this means, says Jensen,\u00a0is it can make recommendations as to what to do and tell you why it has come\u00a0up with its\u00a0conclusions.\u00a0Agentic AI can now understand things like sequences of events. It\u00a0can reason through new problems that may\u00a0demonstrate\u00a0similar\u00a0sequences.\u00a0It\u00a0can reason what will likely happen next.\u00a0It\u00a0can encounter something new upon which\u00a0it has\u00a0not been trained or\u00a0recognizes\u00a0and\u00a0nevertheless\u00a0determine what\u00a0to do.\u00a0<\/p>\n Moreover, Jensen described\u00a0the ability for programmers to easily make customized\u00a0LLMs for specific needs of a certain business\u00a0and then\u00a0have that LLM combine with more generic LLMs with greater\u00a0training\u00a0and more simulations for tailored outputs.<\/p>\n Jensen says the result of all this\u00a0is\u00a0a creative, helpful assistant that can think\u00a0through\u00a0what needs to be done based on what you have done in the past, the\u00a0requirements,\u00a0goals,\u00a0and visions of your business, and\u00a0similar\u00a0outcomes\u00a0across the ecosystem.\u00a0It\u2019s an agent that understands and can interact with our world.<\/p>\n As\u00a0you might expect, Jensen was long on optimism but a little short on what\u00a0all\u00a0has to\u00a0be in place for agentic AI to work as promised.\u00a0It remains to be seen if Jensen and others are right. But there is enough evidence at the show for me to think that\u00a0agentic AI is real. Just how real, we don\u2019t know yet.<\/p>\n But it\u2019s enough of a possibility for legal not to ignore.<\/p>\n For Legal:\u00a0<\/strong>A<\/strong>\u00a0Blessing\u00a0<\/strong><\/p>\n It\u2019s\u00a0easy to write all this off when it comes to legal. The\u00a0conventional\u00a0view is that\u00a0no lawyer in their right mind would let a bot\u00a0unilaterally\u00a0act and make decisions for it. Much too dangerous.<\/p>\n And of course, agents are probably not going to happen in legal anyway.<\/p>\n But as my\u00a0experience\u00a0with Atlas and my car\u00a0demonstrate, agentic AI being an erstwhile companion even in legal may not be that far off.<\/p>\n Such a legal companion could also be a\u00a0tremendous\u00a0aide. It could almost be like a practice mentor, always ready to help.<\/p>\n Here\u2019s an example of how this could work. When I was a young lawyer, I was given a case to practice. I sat\u00a0down\u00a0and drew up a set of\u00a0interrogatories.\u00a0It was not until the eve of trial that I\u00a0learned that out of ignorance, I\u00a0left out a set of\u00a0fundamental, standard interrogatory\u00a0questions: who is your expert and what are they going to say.\u00a0Had I been able to\u00a0merely feed\u00a0the\u00a0initiating\u00a0complaint into an agentic AI tool, it could have run with it and created\u00a0a case playbook that included a\u00a0set of comprehensive\u00a0interrogatory\u00a0questions.<\/p>\n No more digging through countless files to see what others in the firm or elsewhere had done in\u00a0similar\u00a0situations. No more worrying if I missed something critical. No more waking up in the middle of the night wondering if I had filed something in time. Less stress, better results.<\/p>\n Of course, all this depends on the agent giving the why certain things need to be done so the human can decide, given the particular facts of the case,\u00a0that they\u2019re\u00a0appropriate. More importantly, its viability as a tool\u00a0depends on\u00a0a\u00a0human\u00a0being\u00a0in the loop reading, understanding, and evaluating what the agent\u00a0is suggesting.<\/p>\n It\u2019s the human in the loop problem that creates a\u00a0potential curse. The challenge is\u00a0how to avoid it.<\/p>\n And a Potential Curse<\/strong><\/p>\n The truth is it\u2019s\u00a0the human of the loop\u00a0that can\u00a0screw things up.\u00a0For example,\u00a0I had a case where a tragic\u00a0weather-related accident injured\u00a0and killed several people.\u00a0Clearly,\u00a0they had nothing to do with what happened other than they were in the wrong place at the wrong time.\u00a0But one of the lawyers for another defendant in the case\u00a0filed\u00a0their standard set of pleadings which included the claim that the victims negligently contributed to their injuries and deaths. It was wrong\u00a0for multiple reasons.\u00a0Suffice it to say, the media picked it up and the client was\u00a0embarrassed\u00a0and\u00a0possibly\u00a0prejudiced.<\/p>\n I mention that here because it is precisely this human in the loop problem that poses the\u00a0danger\u00a0for\u00a0lawyer use of\u00a0agentic AI.\u00a0It\u2019s the temptation or\u00a0inability\u00a0to\u00a0critically\u00a0think\u00a0through\u00a0a problem and determine if what the agent suggests is appropriate given the situation.\u00a0It\u2019s\u00a0the same human in the loop problem we have with cybersecurity: you can warn and warn to not click on unknown links but sooner or later,\u00a0a human in the loop\u00a0will ignore the warning and do it anyway.<\/p>\n I fear that will end up being the curse of things like agentic AI. It\u2019s\u00a0too easy and tempting to overrely on it,\u00a0particularly\u00a0for busy or, for that matter, lazy lawyers who don\u2019t take the time to treat the\u00a0agent\u2019s\u00a0roadmap with some skepticism. Not to mention the fact, as I have discussed before, GenAI tools and LLMs have the propensity to rot our brains and diminish\u00a0the\u00a0critical\u00a0thinking skill necessary to be discriminating.<\/p>\n The fact that so many lawyers are getting caught citing nonexistent cases proves the point,<\/p>\n The Agentic Future<\/strong><\/p>\n But for those who do and can think critically, I\u00a0can\u00a0see the advantages of agentic AI. It\u2019s like my car. I can tell it where I need to end up and it maps the route and by and large drives itself there. It\u00a0saves\u00a0me time and energy\u00a0and, in many instances, makes it less likely for a mistake to occur. But I don\u2019t go to sleep while\u00a0it\u2019s\u00a0doing so because sometimes it will decide to act in a way not\u00a0appropriate\u00a0for the circumstances\u00a0at hand. That\u2019s when my skills and\u00a0experience\u00a0come into play.\u00a0<\/p>\n The challenge we have as a\u00a0profession\u00a0is to make sure we don\u2019t end up with a profession where those using the agents don\u2019t know how to\u00a0drive.\u00a0That requires attention to training and the challenges. It also requires understanding what agentic can \u2014 and can\u2019t do.\u00a0<\/p>\n One thing we can\u2019t do is ignore it. Things that you say will never happen, indeed don\u2019t happen until they do. And by then, preparing is\u00a0too late.<\/p>\n Stephen Embry is a lawyer, speaker, blogger, and writer. He publishes\u00a0TechLaw Crossroads<\/a>, a blog devoted to the examination of the tension between technology, the law, and the practice of law<\/strong><\/em>.\u00a0<\/p>\n The post CES 2026 And\u00a0Agentic\u00a0AI In Legal: It\u2019s Not Going To Happen \u2014 Until It Does<\/a> appeared first on Above the Law<\/a>.<\/p>\n Conventional\u00a0thinking\u00a0is\u00a0that\u00a0agentic AI is thrown around enough to be\u00a0a\u00a0clich\u00e9.\u00a0We aren\u2019t\u00a0sure what it means.\u00a0We aren\u2019t sure what it does.\u00a0Whatever it is and does really doesn\u2019t mean much for legal, right?<\/p>\n I\u2019ve heard all sorts of\u00a0tech\u00a0concepts thrown around\u00a0during\u00a0the first two days\u00a0of\u00a0CES<\/a>: robotics, longevity, vertical AI, industrial AI. But the one talked about the most is agentic AI and what it can do.\u00a0<\/p>\n I admit\u00a0I\u2019ve been a skeptic.\u00a0But the press\u00a0conferences\u00a0and keynotes by the big players like\u00a0Nvidia<\/a>,\u00a0Samsung<\/a>,\u00a0LG<\/a>,\u00a0and a smaller\u00a0player,\u00a0AGI Inc.<\/a>,\u00a0suggest\u00a0that agentic is not just some pie in the sky\u00a0hyperbole\u00a0but a real thing that can do real things\u00a0for\u00a0real\u00a0people.<\/p>\n As\u00a0Amit Jain<\/a>, CEO of\u00a0Luma\u00a0AI<\/a>,\u00a0put it\u00a0as a guest in\u00a0the AMT\u00a0Keynote, \u201c2026 will be the year of the agents.\u201d<\/p>\n I\u2019m\u00a0now\u00a0convinced that agentic AI is enough of a thing for legal to begin thinking about its benefit and risks.<\/p>\n What\u00a0<\/strong>Agentic\u00a0<\/strong>AI May Mean<\/strong><\/p>\n What all\u00a0these\u00a0keynote\u00a0speakers talked about was the\u00a0potential for\u00a0agentic AI to create an honest to God assistant. Not a dumb GenAI assistant that tries to answer\u00a0just\u00a0what you ask, albeit often\u00a0unsuccessfully. But a real assistant who answers you, makes suggestions for next steps, helps you think\u00a0through\u00a0solutions, and helps you\u00a0implement them. It\u2019s like\u00a0having\u00a0a\u00a0devoted\u00a0friend\u00a0at your side whose only job is to help in every way.\u00a0<\/p>\n Some\u00a0consumer\u00a0examples\u00a0mentioned by\u00a0CES speakers\u00a0demonstrate what\u00a0agentic\u00a0AI\u00a0may\u00a0do\u00a0for\u00a0business.\u00a0An\u00a0agent can decipher what\u2019s in your\u00a0fridge\u00a0and\u00a0create a\u00a0recipe suggestion based on what\u2019s there and what you have historically liked. It can figure out how to season the food based on your taste. It can help you plan an anniversary night out with your spouse based on likes and dislikes it\u00a0knows.\u00a0<\/p>\n It can call you an Uber based on\u00a0your\u00a0request to get you an Uber. And then provide you with all the information you need\u00a0to catch the Uber. No more wandering around airport trying to figure out where and\u00a0how to catch your Uber.<\/p>\n It\u00a0doesn\u2019t take much imagination to think of\u00a0examples\u00a0that may aid\u00a0business and even legal.<\/p>\n I\u2019m sure you\u2019re thinking, as\u00a0I was, it\u2019s the\u00a0CES\u00a0same old, same old. Wild claims backed by extravagant productions with\u00a0heartwarming\u00a0videos about how AI\u00a0is\u00a0going to change our world. But this felt a little different once I reflected on a couple of\u00a0my own\u00a0recent\u00a0experiences\u00a0and\u00a0after\u00a0I heard Jensen Huang<\/a>, founder and CEO of\u00a0Nvidia,\u00a0speak.<\/p>\n
\n![\"\"](\"https:\/\/i0.wp.com\/abovethelaw.com\/wp-content\/uploads\/sites\/4\/2025\/06\/GettyImages-2215793111-300x169.jpg?resize=300%2C169&ssl=1\" "\\"\\"")
<\/figure>\n