1
00:00:05,000 --> 00:00:09,280
Welcome back, our dedicated listeners.

2
00:00:09,280 --> 00:00:14,680
We feel so blessed with the amazing response we have been getting for this podcast.

3
00:00:14,680 --> 00:00:17,440
Okay, okay, enough self praise now.

4
00:00:17,440 --> 00:00:23,640
Let me get back to another episode of potential award winning podcast.

5
00:00:23,640 --> 00:00:28,320
Lexdiscuss Cafe with huge emphasis on the word potential.

6
00:00:28,320 --> 00:00:34,360
This is the second part of our conversation with Mr. Biswarup Ghosh, an engineer turned

7
00:00:34,360 --> 00:00:38,920
lawyer hailing from the prestigious IIT Kharagpur.

8
00:00:38,920 --> 00:00:45,400
In the first part of this podcast, Biswarup shares his experiences of studying at IIT

9
00:00:45,400 --> 00:00:50,440
Kharagpur and transitioning into a legal role at a tech company.

10
00:00:50,440 --> 00:00:57,340
He also highlights the benefit of his alma mater, including the exposure to various disciplines

11
00:00:57,340 --> 00:01:00,060
and world class education.

12
00:01:00,060 --> 00:01:06,260
He also discusses his professional journey from IP law to contracting and privacy and

13
00:01:06,260 --> 00:01:09,500
his current role at Cambian networks.

14
00:01:09,500 --> 00:01:16,100
He also touches on the importance of lawyers in tech companies adapting to the environment

15
00:01:16,100 --> 00:01:19,980
and communicating effectively with engineers.

16
00:01:19,980 --> 00:01:25,180
Our conversation then shift to the challenges of keeping up with the regulatory changes

17
00:01:25,180 --> 00:01:29,500
and open source software in the context of law and technology.

18
00:01:29,500 --> 00:01:37,100
Lastly, Biswarup emphasizes the importance of both legal and business knowledge for success

19
00:01:37,100 --> 00:01:43,060
in the field and the need for law schools to teach practical skills such as policy implementation

20
00:01:43,060 --> 00:01:44,620
and technology addressing.

21
00:01:44,620 --> 00:01:51,780
Today, we continue our conversation unraveling more layers of his experiences, challenges

22
00:01:51,780 --> 00:01:53,780
and unique perspectives.

23
00:01:54,620 --> 00:02:00,300
So Biswarup, the next question is really important and you know the buzzword these

24
00:02:00,300 --> 00:02:06,340
days, what do you think generative AI is going to do to the legal profession and what legal

25
00:02:06,340 --> 00:02:09,900
profession should be doing in this regard?

26
00:02:09,900 --> 00:02:12,940
I was just hoping this will come up.

27
00:02:12,940 --> 00:02:13,940
So I'm happy.

28
00:02:13,940 --> 00:02:19,880
Today we have two crowds, one is who have heard of generative AI and another have seen

29
00:02:19,880 --> 00:02:21,140
and done that.

30
00:02:21,140 --> 00:02:26,620
If I were to speak to set up people who have heard about generative AI but have not seen

31
00:02:26,620 --> 00:02:32,380
the magic or have not had their hands on that particular thing, I would get a response that

32
00:02:32,380 --> 00:02:33,980
this is again a fad.

33
00:02:33,980 --> 00:02:39,180
But myself being on that other group where I have actually seen, I have put my hands

34
00:02:39,180 --> 00:02:45,060
on and have used on my personal capacity, generative AI, the GPTs of this world in fact

35
00:02:45,060 --> 00:02:49,820
have gone ahead and worked on some small projects on training of the LLMs, etc.

36
00:02:49,820 --> 00:02:52,260
I can say that this is path-breaking.

37
00:02:52,260 --> 00:02:53,820
Let me get into the specifics.

38
00:02:53,820 --> 00:02:58,900
Now generative AI, I mean we end up speaking a lot about generative AI, let me take a step

39
00:02:58,900 --> 00:03:04,500
back and let me speak about AI itself, because generative AI is essentially when we are mostly

40
00:03:04,500 --> 00:03:10,020
talking about prompt-based, text-to-text kind of a generation sort of thing.

41
00:03:10,020 --> 00:03:13,220
But AI in itself is extremely powerful.

42
00:03:13,220 --> 00:03:17,980
I have to break down my work and break it down into what kind of technologies today

43
00:03:17,980 --> 00:03:19,400
intervene in my world.

44
00:03:19,400 --> 00:03:25,160
So one part of it is I get tons of emails and lot of it is something I have to bucket

45
00:03:25,160 --> 00:03:30,100
between what I do within 5 seconds, what I essentially need to put in some time and collect

46
00:03:30,100 --> 00:03:34,600
certain information and then reply within certain period of time, say a day, and what

47
00:03:34,600 --> 00:03:40,280
is a little spread out, I mean I can actually reply back in say maybe a week and it's fine.

48
00:03:40,280 --> 00:03:44,900
Now generative AI is something which helps me in prioritizing this task, much better

49
00:03:44,900 --> 00:03:49,780
than how I would otherwise do it, where I go into outlook and create tasks for myself.

50
00:03:49,780 --> 00:03:53,980
So I keep on creating tasks, never being able to complete those.

51
00:03:53,980 --> 00:03:57,860
But with these kind of Microsoft insights, if you have seen these coming up and saying

52
00:03:57,860 --> 00:04:02,580
that hey, you have a mail which you received on say 3 days back and this is waiting for

53
00:04:02,580 --> 00:04:08,020
your reply and then this is something which are the kind of documents which go with respect

54
00:04:08,020 --> 00:04:11,860
to you being able to reply, well, fantastic use case.

55
00:04:11,860 --> 00:04:16,820
The second bit is I need to then bucket it into the kind of case matters that I need

56
00:04:16,820 --> 00:04:17,820
to handle.

57
00:04:17,820 --> 00:04:22,020
Now here, generally, I have to prep, going back to this open source example, some open

58
00:04:22,020 --> 00:04:25,860
source software comes to my table, I have to go ahead and do the review of the license

59
00:04:25,860 --> 00:04:31,060
and then go back and give an approval or some suggested steps to take care of this, right?

60
00:04:31,060 --> 00:04:35,860
Now again, AI is extremely helpful in number 1, if I were to go ahead and figure out what

61
00:04:35,860 --> 00:04:40,440
those open source software are, these new software which are today developing which

62
00:04:40,440 --> 00:04:46,440
can pick it up, utilizing AI and almost anticipate what are those kind of code snippets, fantastic.

63
00:04:46,440 --> 00:04:51,560
The workflow pieces that are getting built up in this area to manage the workflows that

64
00:04:51,560 --> 00:04:57,140
I push for someone's recommendation, for someone's approval, I receive certain approval or certain

65
00:04:57,140 --> 00:05:02,520
comments from them, I collate all these together and send back as an advice to my internal

66
00:05:02,520 --> 00:05:05,880
customer, Generative AI has a fantastic use case.

67
00:05:05,880 --> 00:05:10,360
Going to the pieces where I need to do drafting, I need to draft a contract, I need to draft

68
00:05:10,360 --> 00:05:15,960
say a patent application, I need to draft any of the applications which are non-contentious.

69
00:05:15,960 --> 00:05:19,240
Generative AI gives a fantastic first step.

70
00:05:19,240 --> 00:05:24,960
Generative AI alters when we want to create a work product which is a final workflow.

71
00:05:24,960 --> 00:05:31,440
I am continuously seeing today that Generative AI is not there, AI is not there to basically

72
00:05:31,440 --> 00:05:33,840
come up with such kind of work product.

73
00:05:33,840 --> 00:05:39,640
That's where an AI enabled lawyer is a much better place because it frees me up from a

74
00:05:39,800 --> 00:05:43,880
lot of mundane kind of a task and helps me focus on two things which are essentially

75
00:05:43,880 --> 00:05:46,440
material for me as well as the company.

76
00:05:46,440 --> 00:05:52,480
The last piece where it comes down to the contentious issue and this is where specifically

77
00:05:52,480 --> 00:05:59,800
for me, I am not super excited about utilizing Generative AI when say a code application

78
00:05:59,800 --> 00:06:05,440
has to be drafted or where there is some sort of these kind of applications which essentially

79
00:06:05,440 --> 00:06:11,080
have to be filed in front of courts in India or in most of the countries in APAC are concerned.

80
00:06:11,080 --> 00:06:15,880
That I am not yet very excited about, that is not something which I can see that is happening

81
00:06:15,880 --> 00:06:17,720
to that kind of a scale.

82
00:06:17,720 --> 00:06:22,720
So a long-winded way of saying this, being in the position where I have already utilized,

83
00:06:22,720 --> 00:06:27,440
I have used certain part of it, I can see that this technology is definitely something

84
00:06:27,440 --> 00:06:29,960
which is path-breaking, no doubt about it.

85
00:06:29,960 --> 00:06:36,400
Where we utilize, how we utilize, some places has got immediate applications, some places

86
00:06:36,400 --> 00:06:40,640
I am not sure sure, maybe down the line it is going to come up and of course there are

87
00:06:40,640 --> 00:06:43,720
some other places where it doesn't make sense to touch it.

88
00:06:43,720 --> 00:06:48,880
Now one last piece is about the research, we end up continuously required to do research

89
00:06:48,880 --> 00:06:54,480
on what are these kind of laws, what are these kind of implications, regulation, order, notices, etc.

90
00:06:54,480 --> 00:06:59,840
I mean that's the fantastic use case of AI and I have been utilizing in and out AI for that.

91
00:06:59,840 --> 00:07:05,400
Being able to type it down in natural language and being able to get a response which is

92
00:07:05,400 --> 00:07:12,120
much better than a Google search result, this is my type immensely, so that's what it is.

93
00:07:12,120 --> 00:07:16,600
And your second part of the question about what should lawyers do about it, I mean there

94
00:07:16,600 --> 00:07:23,400
is nothing stopping the lawyers from starting to try, I mean see this as a tool and then

95
00:07:23,400 --> 00:07:29,160
also keeping in mind that the more we remain inhibited with something new, the chances

96
00:07:29,280 --> 00:07:34,320
are that you know the wave is only going to bite us, so therefore I mean what I have been

97
00:07:34,320 --> 00:07:38,720
doing, I can't really go ahead and represent the whole community here but what I have been

98
00:07:38,720 --> 00:07:43,120
doing is anytime there is a new tool, I have been trying to make use of it, figuring out

99
00:07:43,120 --> 00:07:47,520
where it makes sense, trying to understand what sort of risk it carries while we are

100
00:07:47,520 --> 00:07:52,480
doing this and then wait for the right opportune moment to actually put it into work.

101
00:07:53,440 --> 00:07:58,240
Completely agree with your thoughts there but you know one of the reason that we invited

102
00:07:58,240 --> 00:08:03,120
you for the podcast because of having a technical background earlier and I am sure you have

103
00:08:03,120 --> 00:08:08,320
met people who don't share that background, why do you think there is so much reluctance

104
00:08:08,320 --> 00:08:13,840
in adopting new technology among lawyers and my follow up question to that would be considering

105
00:08:13,840 --> 00:08:20,480
that you are in a tech company where everybody is adopting AI left right center, so do you

106
00:08:20,480 --> 00:08:25,040
think the legal department also gets influenced a little more to do that or do you believe

107
00:08:25,040 --> 00:08:25,840
that's not the case?

108
00:08:26,800 --> 00:08:31,760
Well Abhi, you have to see that there are these different worlds which even exist already.

109
00:08:31,760 --> 00:08:37,280
This is a personal story, so I have been to a district court for one of our matters

110
00:08:37,280 --> 00:08:41,920
going on there and there I see what's the process, let me explain to you the process.

111
00:08:41,920 --> 00:08:48,640
There is a lawyer who is going to sit down and first write the application on their hand,

112
00:08:48,640 --> 00:08:53,920
so when I sit in front of them, I do not see a laptop in front of them, our handheld devices

113
00:08:53,920 --> 00:08:59,680
are in fact non-smart phones right so to say the dumb phones and then the process moves on

114
00:08:59,680 --> 00:09:04,480
so that lawyer would draft it and then they would send it to someone who is going to again write it

115
00:09:04,480 --> 00:09:09,360
and mind you I mean most of our district courts for their own good they work in the native language

116
00:09:09,360 --> 00:09:15,440
I mean they will work in say Hindi or Bengali or say Telugu or Tamil etc. So now this is a process

117
00:09:15,440 --> 00:09:21,600
which gets followed essentially if we are trying to say today that why is a lawyer who is today

118
00:09:21,600 --> 00:09:26,720
writing it on the paper and then getting it typed on say a computer or maybe a typewriter

119
00:09:26,720 --> 00:09:30,800
I just saw a typewriter a few days back so I am telling it with a lot of conviction that this

120
00:09:30,800 --> 00:09:37,440
happens why is it that guy or that person not inclined to use AI so we are not really asking

121
00:09:37,440 --> 00:09:43,440
the right question because you see none of us actually reached our graduation before having

122
00:09:43,440 --> 00:09:48,160
completed our first class on the other side you are looking at large law firms you are looking

123
00:09:48,160 --> 00:09:53,520
at large corporates and I am going to only speak about the corporates over here now in corporates

124
00:09:53,520 --> 00:09:58,880
we end up seeing that our peers our colleagues end up using AI or I mean even before AI these

125
00:09:58,880 --> 00:10:04,000
kind of automation tools this kind of SaaS software left right and center so in that case we are not

126
00:10:04,000 --> 00:10:09,920
just influenced or I would say we are essentially expected to use this software as efficiently if

127
00:10:09,920 --> 00:10:16,000
not more as our other colleagues in the department so therefore the adoption of these technologies

128
00:10:16,000 --> 00:10:21,040
are naturally much higher in places in this kind of corporate legal department

129
00:10:21,040 --> 00:10:27,040
so even today I mean most of the corporate legal departments would have one or other form of case

130
00:10:27,040 --> 00:10:32,240
docketing system or a contract management system these are already there they would likely to have

131
00:10:32,240 --> 00:10:37,360
an e-signature tool if there are multinational contracts which are being executed so these are

132
00:10:37,360 --> 00:10:44,080
already there AI has a layer on top of it so therefore it's not too difficult for any of

133
00:10:44,080 --> 00:10:50,000
these lawyers to promote themselves into using this that's what my thesis is that if we are

134
00:10:50,000 --> 00:10:56,480
expecting the larger chunk of lawyers within India to today sit down and start using it I don't think

135
00:10:56,480 --> 00:11:01,600
that's the correct expectation on the other side if we are expecting the corporate legal departments

136
00:11:01,600 --> 00:11:07,680
to use as much as they can yes that expectation is absolutely correct and that's what is happening to

137
00:11:07,680 --> 00:11:13,920
it wow great insight said so as we are running out of time the last question I'll ask you

138
00:11:14,560 --> 00:11:21,600
so what would your advice be to young lawyers who are graduating in the era of generative AI

139
00:11:22,640 --> 00:11:25,120
well that makes me sound very old but

140
00:11:25,360 --> 00:11:33,200
but you're not that old I'd say younger lawyers right younger lawyers

141
00:11:37,840 --> 00:11:44,960
yes I mean it's imperative to basically on two sides to know this one is how to use it and also

142
00:11:44,960 --> 00:11:50,720
keep our minds open I mean this is a clear shot of a message to any of the younger folks who are

143
00:11:50,720 --> 00:11:56,080
coming out of law schools and getting into the profession so at one point or the other you are

144
00:11:56,080 --> 00:12:01,200
going to get asked what are the kind of risks which are involved if a team of yours or a client

145
00:12:01,200 --> 00:12:07,520
of yours has to either develop an AI or has to introduce AI into their products it's almost like

146
00:12:07,520 --> 00:12:12,320
earlier we used to say every company is a technology company today every company is an AI

147
00:12:12,320 --> 00:12:17,200
company either they're using it or they are essentially making it so therefore the lawyers

148
00:12:17,200 --> 00:12:23,280
have to be prepared on two areas first is how to use it unless we ourselves are hands-on with

149
00:12:23,280 --> 00:12:28,720
respect to this it's difficult to get a feel get an understanding get a flavor of how does it work

150
00:12:28,720 --> 00:12:33,040
right I mean if I would not have ever used a stability diffusion I do not know what are

151
00:12:33,040 --> 00:12:37,840
the kind of distortions which come out of such kind of the software if I were to never use a

152
00:12:37,840 --> 00:12:42,720
chat gpt I do not know what are the kind of extent and where it can start hallucinating

153
00:12:42,720 --> 00:12:47,120
so that is one so we need to start using and we need to start understanding that

154
00:12:47,120 --> 00:12:51,920
as much as we can and on the other end we also need to keep our eyes and ears open as to what

155
00:12:51,920 --> 00:12:57,120
are the kind of risks which are there I don't really want to make it a session of you know

156
00:12:57,120 --> 00:13:04,080
describing the risks but if we were to listen in closely talk to people who are working with it or

157
00:13:04,080 --> 00:13:09,760
developing it or you know who are actually using it we will immediately start to understand that

158
00:13:09,760 --> 00:13:16,560
it starts creating the risk in certain areas of these privacy in the areas of in fact you know

159
00:13:16,560 --> 00:13:22,480
ethics so on and so forth so these two are the specific things that I would like someone to

160
00:13:22,480 --> 00:13:28,000
focus and if they do they can stay assured that when they are facing a situation where they have

161
00:13:28,000 --> 00:13:34,400
to either take a call that should I buy an AI tool they are best players or they have to take a call

162
00:13:34,400 --> 00:13:39,600
that yes this organization wants to introduce an AI software they are best players to give

163
00:13:39,600 --> 00:13:45,920
them an insight great insights Vishwaroop so Vishwaroop thanks a lot for joining this podcast

164
00:13:45,920 --> 00:13:51,200
trust me it was really amazing talking to you thank you very much Avi it was fantastic talking

165
00:13:51,200 --> 00:14:00,800
to you as well as we sign off remember that law and technology are kindred spirits they coalesce

166
00:14:00,800 --> 00:14:08,000
innovate and redefine boundaries so dear listeners keep your legal compass calibrated

167
00:14:08,000 --> 00:14:15,680
your tech binoculars focused and your curiosity unbridled thank you for joining us on this

168
00:14:15,680 --> 00:14:26,400
odyssey subscribe share and keep the conversation alive stay curious stay caffeinated