00:00 The University of the Philippines has drafted guidelines on the responsible use of artificial
00:04 intelligence amid the technology's rapid advancement.
00:08 UPAI Program Assistant Professor Miguel Francisco Ramolona joins us now via Zoom.
00:14 Professor, thank you so much for joining us.
00:16 You've read this draft of guiding principles on the responsible use of artificial intelligence.
00:23 What do you think of this draft so far?
00:28 So far, we think that there's potential in the draft.
00:33 There's a lot of topics there that are already covered, especially with regards to the ethics
00:39 of the use of artificial intelligence.
00:42 But we do not want this policy or this draft of this policy to be a basis for hindering
00:51 the development of AI.
00:53 So what we want is to have a more democratic approach in drafting these policies.
01:00 That's why these policies were revealed to the public before being finalized.
01:06 Because of the responses and because of the perception of people, especially with regards
01:16 to the negative sides of AI, we also wanted to make sure that this doesn't send the wrong
01:22 message to everyone, that we still want to develop and still use AI, that we are not
01:29 actually limiting its use, but we're just looking at what are the possible repercussions
01:34 and what are the possible ethical situations.
01:39 So for now, Professor, the principles in this draft do not include specific policies on
01:47 the dos and the don'ts in the use of AI.
01:51 What's with it?
01:52 Well, for now, I think it's better to have what is known as a soft regulation when it
01:59 comes to these types of emerging technologies.
02:04 Because if we put in hard regulations, then we're limiting ourselves and we might be actually
02:10 hurting instead of helping both our students and our teachers.
02:15 This tool might be helpful in terms of expanding the range of what is currently easily in reach
02:23 of the students, like trying to access newer information might be harder, even though with
02:29 the existence of Google and Wikipedia, that's constantly updating.
02:33 When it comes to state of the art, it's harder to find that.
02:37 And ChatGPT might be a gateway that will have them access this information.
02:41 You mentioned ChatGPT, Professor, you mentioned ChatGPT, and many are already worried that
02:47 ChatGPT opens the door for cheating and plagiarism.
02:51 How imperative is it for schools to promote the positive use and mitigate the negatives
02:57 of AI?
02:59 So right now, in terms of plagiarism and for cheating, for plagiarism, there's actually
03:07 tools right now that help us identify AI writing as opposed to human writing.
03:13 Because AI has a certain style of writing, which combines a lot of, which is essentially
03:19 algorithmic based.
03:22 So sorry, I heard an echo.
03:25 But in any case, so with regards to using this AI writing detection tool, what we use,
03:33 at least in UP Diliman, is Turnitin.
03:35 So Turnitin has an AI detection tool.
03:38 We are currently testing the limits of this AI detection tool, trying to see how far it
03:45 can detect these AI writings, what prompts would be possible, or if I edit the text,
03:52 would it still be able to detect it?
03:55 And so far, at least in our tests so far, we were able to still detect a lot of AI writing,
04:03 even though there's been some minor tweaks in the wording or in the sentence construction,
04:09 but it's still able to detect it as AI writing.
04:12 So if you're worried about that, that's one way to mitigate it.
04:17 So do you think that other universities, professor in the country should also make similar efforts,
04:22 just like what UP Diliman has did?
04:27 I think efforts into trying to identify the perception, especially of faculty and students,
04:33 should be done in the entire country.
04:36 In fact, maybe not just in education, but also in affected industries.
04:41 Just the perception and how they feel that this would impact their lives.
04:45 Would it make it easier?
04:46 Would it make it harder?
04:47 Is it a threat to them?
04:49 Because these sorts of things is what helps in the development of the technology.
04:54 So if we have tools, if we have a more proper understanding of at least the implications
05:01 in terms of our lifestyles, then it's a little bit better to, a little bit easier for us
05:10 to work around the threats, to work around the hazards that are involved.
05:17 Because right now, this is a fast moving technology.
05:20 We can't stop it.
05:22 What we can do is adapt to it in terms of education.
05:25 That's what we're trying to do.
05:26 We're not trying to hinder its growth.
05:28 We're trying to use it.
05:30 We're trying to ensure that it's not abused.
05:34 And that we can identify, or we can just, instead of having essays that are submitted
05:41 or ChatGPT create essays in place of essays that are required, we can ask them to make
05:47 two essays, one that ChatGPT wrote and the other one being a modification of what ChatGPT
05:54 wrote.
05:55 In this case, they're learning in more dimensions than what they can just by reading.
06:00 Okay.
06:01 So, Professor, Professor, I'm sorry, we've run out of time.
06:04 Thank you so much for explaining to us this new draft guidelines.
06:09 And hopefully we will have the final version sooner rather than later.
06:13 UP AI Program Assistant Professor Miguel Francisco Ramolona.
