AI technology as it stands is not even close to actually replacing a human lawyer though.
Maybe it could "augment" the lawyer's human intelligence, but as it stands, we would need real breakthroughs in the field of AI to actually do logical reasoning (and I'm not talking about black-and-white hardcoded Prolog).
If anything, contemporary AI tech could help perform diagnoses in the medical industry. Indeed, there are several deep learning startups tackling this problem. Diagnosing medical conditions is essentially a classification problem so you just need the right training data and a convnet.
Computer searching has already removed the need for the huge historical cases where a large group of lawyers and pa's would search through old emails and docs to see if they could find something about a big case. Now that's generally done by computer searching. You can call that AI if you have a need to, but the trend is clear - white color work that is easily automated will become more efficient, leading to fewer workers in general.
Searching isn't AI, and one data point doesn't establish a trend. The very first steps to replacing a human lawyer with an AI are still very much unsolved research problems in the field of computing.
4 comments
[ 8.5 ms ] story [ 23.4 ms ] threadThe ability to refer to any. case. ever. to provide a potentially watertight argument will be pretty hard to beat.
Maybe it could "augment" the lawyer's human intelligence, but as it stands, we would need real breakthroughs in the field of AI to actually do logical reasoning (and I'm not talking about black-and-white hardcoded Prolog).
If anything, contemporary AI tech could help perform diagnoses in the medical industry. Indeed, there are several deep learning startups tackling this problem. Diagnosing medical conditions is essentially a classification problem so you just need the right training data and a convnet.