Racist Robots Are Replacing Lawyers | The Daily Show With Trevor Noah 

Somebody needs to put him in a movie. Dramatic. The timing thing is dynomite

If they ever fired Trevor, Chang for the host

Jay Be he is hilarious! Especially when he gets mad

I suggest stephen colbert.

😂😂so true

I have a feeling they know what to expect going in. I doubt that they're conned into thinking that it will be a standard, non-comedic interview.

Yeah

True😂

Very true...

Zambia too

hahahaha i found out the hard way

You're right 😂😂😂

Honorable "Judge Dredd" Alexa

"My existence, well grotesque and incomprehensible to you, saves life's" but with a cold, robotic voice was the best part IMHO

can't handle the doot

@@johannestonnies7898 Can't handle the suit

True

Another problem is that in law you shouldn't analyze all the data in all the cases with the same standards. Yes, the law is the same for all, but applying it to the particular case with the same logic everytime leads to unfair results, because the law, as a human creation, is unperfect and can't provide a good solution to every imaginable hypothesis.

True

@j4g0 I think you misunderstand how current state of the art AI - neural networks - work. Programmers construct blank network with no knowledge. Then they feed it data and it learns, for example, to classify it. Facebook tried that with chatbot and they had to turn it off because it became... racist. This does not mean that programmers put implicit bias into code. This means that data - chats of facebook users - was biased, i.e, racist.

Like people then - they have a hard time recognising and overcoming their prejudices too.

Yes that is very likely indeed and it might already be the case for the current level in AI - The frame of which the AI is set to operate within is determined by the Human(s) that has created the AI - I you are a Hacker or do other cybercrime, your AI could very well be set to work within this parameter or frame = Criminal AI or if it has a body of some sort then yes a criminal robot.

well hackers are developing AI self-growing virus

@Meijer Nelly Surely that is the bigger gamble of the whole AI development - They expect that AI´s will make huge scientifical steps and solutions that it would otherwise take intellectual and scientific groups years to accomplish - But it is never the less a gamble - exactly as you say - AI´s might create their own languages and has already done that. A hugely powerful AI with full connection to the net could essentially infect a whole lot of computers and systems with a "virus" in a language we don't know.

JoshAndBooze That sounds pretty conspiracy theory-esque to me. The AI's simply drawing a conclusion based on the described action of person X. The context, and how that correlates to what is described in the law as illegal. If the descsription does not fit, then the AI will conclude the person to be not guilty & can demonstrate why.

JoshAndBooze I'm not the best at computers but im 98.76% sure thats not how it works. And as for the programmers being emotional part: you do realise they're probably just designing a system to search for markers that increase the risk and that those markers will be decided based on currently available scientiffic data. The programmers would have to intentionally ignore facts in order for their emotions to be a problem.

The problem of AI lawyers wouldn't be them having emotions, but the lack of emotions. The abilities required to apply law include empathy, emotional intelligence, intuition and more. Lawyers need emotions, just like any other humanities professional.

Khalkara, the thing is that knowing what the law says (or giving the law a meaning, really) and seeing if a situation fits into a legal description, is not a mechanical process solved with some kind of mathematical logic. That is the hard stuff. Doing that is basically all what the legal profession is about.

Andrés Falcone Can you elaborate? Because the way I see it, if you had a unfairly received a parking ticket because you parked on the wrong spot supposedly, say in front of a certain building. And whilst the law says you indeed can't park in front of such a building UNLESS it is past 18:00, assuming you would explain to the lawyer AI that you were parked there from 18:30 to 19:00 or whatever then the AI should be able to make an objective case that this person is not guilty based on that. I don't see where having emotion becomes a positive/negative, especially in regards to a profession that is supposed to be about objectivity.

calm yourselves the only thing robots will take over is physical labour. The rest is complicated for the so called AI.

Discounting retail and other frontline service jobs(which take up like half of the US workforce), Most physical labour jobs are safe from automation. There are serious engineering and cost issues that prevent today's line cooks, assembly workers, firefighters, janitors and lumberjacks from being replaced completely by robots. And we've mostly reached that plateau already. No venture capitalist is going to invest billions designing an ultra high tech logger to replace cheap lumberjacks. Meanwhile replacing white collar jobs is just a software challenge which the massive computing industry is going to work on anyway.

How can you replace a teacher with a robot?

Its happening in investment industry, all of those investment managers being replaced by complicated algorithm

i’ll be dead by then

Dragold Anime That doesn't mean that the potential of these AI should be ignored. They can search through thousands of legal documents in minutes saving precious time and money on the logistical/clerical side of legal council. This can lead to much fairer trials as it means less mistakes will be made when it comes to by the book type stuff and things like precedent.

That's better. If there's evidence against you, maybe you aren't innocent. But I do reckon these robots do know what is in your favour, so, I don't know.

Kenny Martin see CGP Grey's video "humans need not apply". It has some great stuff about this.

Kenny Martin I Guess you right on that fact but what if your client is not innocent enough, you still need to produce “legal” documents to win his case. That is what a lawyer is paid to do.

That's exactly what I was thinking. A machine cannot look at the intricacies of the law and suggest loopholes for you to escape out of like a real lawyer can. This 'robot' is basically just a glorified search engine for laws. I am sure that is not very hard to create.

That is a very tricky question - For pure logic there is no merci, mathematic dose not include Merci or similar concepts. On the other hand the AI´s will be set up by humans to being with and what they define as the AI´s frame or human relations parameters will determine if such a concept as Merci is to be included or if it at all can be included. Merci is based on empathic understanding - like forgivenes. To forgive someone means that they essentially have done something wrong or caused harm to another person - If the person deserves merci will depend on circumstances - These can be related to moment of when the wrong was done and/or they can related to a persons history. If an AI will be capable to weigh in such a narrow balanced decision is a very good question. Some parameters will be ""fairly easy"" to set up - like if we are talking about young kids, an AI can learn most of the difference between a child and a grown up because of knowledge database and age. But e.g if a 16 year old teenager goes rouge with an AR15 and shoots his or her classmates and at the same it is found out that the teenager have been severely abused at home, thus being bullied extremely by the class mates, then it all starts to become a very difficult balance act to judge. Can an AI find the right judgement for such a scenario ?

Dude no thats the reason that they use robots, so no judge will have mercy or the opposite.

No, but it will have thanks.

Robots cannot be racist.... Robots have no emotions.. Humans suffer from emotions of different flavours and colours... I may be wrong..... You maybe right..... Let us work with love peace and happiness for all.....

We've already got a soulless uncanny-valley failure who makes unpredictable decisions based on logic we don't fully comprehend and probably needs a couple decades more work before he can be considered to have human-level intelligence.

Bender from Futurama would be a better president..

May be?

i hate you too

I'm a software engineer and I know plumbers and mechanics who make more than I do.

But less bias than humans. It can be made more efficient with better data and adopted learning model.

So you are saying that a lawyer needs to evaluate massive quantities of data, apply context and variables to produce a detailed analysis? Then apply the results of the analysis in an adversarial framework? Yep all skills no AI will never learn!

If ai can replace doctors, I don't see why it can not replace lawyers.

Actually not. The only massive amount of data that lawyers usually handle is jurisprudence, and there's plenty of software to help with that. You usually work with small amounts of data. A short phrase with multiple possible interpretations. The balance of principles that reflect values (and by value I don't mean a number). A judgment where you have to put yourself in the place of another person, to see how a reasonable person would react, or to understand their interests (even emotional interests). A LOT of human interaction. A lot of thought with philosophical basis. We're probably not far from AIs that can do most of those operations, but the results would rely in the values a programmer from his computer thought were more important, that is, a pre-judgment, and not in a case-by-case analysis and weighing of those values. The results would be considered weird and unfair by us humans, and the society wouldn't accept such system as legitimate. You know, in the XIX century there was this idea that judges were mere operators that just applied to the particular case general solutions predetermined by the law, just like robots... but that idea had a shocking fail when faced with the reality that law by itself isn't enough to solve a case, there are contradictions, voids and value judgements everywhere: you need wisdom, good judgment or whatever you call it, a quality that only humans have, where intuition, empathy and emotional intelligence play a big role.

1. Doctors deal with matters of natural science, where everything works with mathematical logic. Everything is correct or false, every phenomenon has a certain result. All is predetermined by the rules of nature. When this kind of reasoning leads to mistakes, it's because a certain rule of nature hasn't been discovered yet. Machines can handle that, provided that the correct rules of nature and the facts can be reduced to numbers in an objective process (easier said than done). Lawyers deal with law, which is a social construct. Even though we know all the rules (we created them, after all), they provide no certain results. Everything can be analyzed from another point of view. A single phrase or situation may have multiple valid interpretations from a purely logic point of view. Values are of the order of the day. Machines can't handle that. 2. AI hasn't replaced doctors, it has become a tool for them (and for lawyers too.) 3. Doctors take important ethical decisions and have a lot of human interaction. Health is not something purely physical. Machines can't deal with that.

Andrés Falcone lawyers rely on previous supreme court rulings or same case scenarios. They rely on what is the written law. It would be easier to ai legal system than medicine. Law is supposed to be blind, meaning that if you committed an offense, taking into account the mitigating factors, a person should get the same punishment as a person who did the same offense of the same circumstances. Unfortunately, law as it is today, the poor gets harsher sentence than a rich guy who can avail of good lawyers. The USA has 1 % incarceration rate, meaning there are 30+ million inmates at any given time. That is twice the number of the runner up. I am sure it is a broken system.

Toriqul Islam Shimanto Seeing as human lawyers are already racist and biased, I'd say it's just going horizontally

The machine isn’t actually racist, it’s running risk assessment code, we are moving forwards towards actual fairness and logic.