Hear professor's prediction on the future of AI tools

ChatGPT, an AI-powered chatbot that interacts with users and can provide lengthy and thorough responses to questions and prompts, is stunning users. Professor Scott Galloway from NYU Stern School of Business joins CNN's Anderson Cooper to discuss. #CNN #News

[Wooster]

Student: chatgpt, write an essay for me. Teacher: chatgpt, grade this essay for me.

[ThePeedz]

I highly respect Professor Galloway, as a Buisness Professor. I wouldn't under any circumstances consider him a tech expert in anything other than running a Tech Business. I am so happy that he pointed that out on the air. Mad respect.

[c.m.]

I put in a few opening lines of a scientific paper on biodiversity into one of those tools, and it ended up writing an abstract of what COULD have been writen in this paper. It pretty much provided a study system, a population, claimed that we collected genetic data of 113 individuals, and claimed we found statistically significant SPECIFIC results.
It was really strange to read. It just made up a study that sounded valid.

So, when the guest here says it can be used to come up with credible sounding propaganda and fake studies… you better believe it!

[steves]

Teachers should be able to ask it if it had written the following text, and it should come back with, “yeah I did that for Johnny yesterday”

[aruranganeshan]

I've observed that individuals who have completed PhDs and MBAs may have concerns about the potential for their expertise to be replaced by technology such as ChatGPT. Similarly, there may be concerns about the impact of technology on the medical field and the potential for doctors' work to be taken over by advanced systems like ChatGPT 4.

[murraymadness]

I worked on natural language AI in grad school and this is what we thought we could do someday. What I did was more like AskJeeves, much simpler, but back then CPU's were running at 33mhz with 8mb of memory.

[vvolfbelorven]

I get the core consequences of AI/AGI:

1) We will not be the source of production. Instead, we will be the bridges - the ones making the AI connect the dots. I think horizontal thinking, creativity, and exploration will be paramount to leverage AI capabilities in almost every area of life.
2) Our little jobs in the knowledge economy (from data entry to executive role) and others that will be impacted by this technology are unimportant in the scheme of the really big things. The implementation of AI definitely means ' job displacements' - mine in database in the regulatory part of the pharmaceutical industry included - but at the same time it means solutions to climate change, energy crisis, and pollution (what to do with plastic?). The only path forward is to reskill.
3) The social aspect of AI has not been discussed yet but I think it has profound implications. Using ChatGPT, I can clearly recall its output from questions I asked. The output it provided was not only useful (as it is in most instances), it was also personalized . This concept, of having a personalized exchange, not only is different from the cold million hits of the search results we get in Google, it also has the major implication of a 'staying' power; meaning we may prefer to chat with the AI AND be more receptive to its output (thoughts?) and defer to it. For instance, we used to say "according to Google, " but moving forward it will likely be "according to ChatGPT..." I think this is where the control part of the technology starts.

As every type of technology, the effects of AI will depend on how we use it. Trains were invented to move large amounts of goods and people; the Nazis used trains to haul people off to the Holocaust.

Disclaimer: I did not use ChatGPT for this comment. I don't think it could have come up with these observations. There may still be some hope left! =D

[latetotheparty]

I was a Freshman in college 1974. I took a computer class, there was a HP5000E that we had to dial up from the computer lab on dedicated lines. In the class I was introduced to Basic, and the teacher had written the textbook because there were so few textbooks introducing computers at that time. There was a program we use use that suggested the computer was a psychological therapist. The dialogue and responses were printed out. I found the program was self-referential and fascinating. I wish I still had the printouts.

[Dryvlyne]

The professor's concern about how AI technology could eventually displace large segments of workers is spot-on. It's going to be critically important for people to stay up to date on the latest advancements and how to use them. As a simple example, I'm still quite alarmed at the number of workers at my employer that don't have a good understanding of how to use Excel. Eventually such disparities in skills will lead to many people getting left behind in job opportunities.

[kurtdobson]

Current AI reflects the same biases and errors present in the data used to train it. A major challenge is how to remove bias and ensure AI is based on 'Ground Truth'. Certainly not the case at present, and I expect there will be a battle over who guides the future development of these AI engines.

[ctrl-shift-run]

Definitely agree with him about how this is a tool that will create economic opportunity. Like if you couldn't code a scraper or a trading algorithm, you can now do that.

That being said, there will still be a need for programmers to put unclear requirements into code, stitch it together, test it and also to ask the question of whether the whole thing makes sense or not.

[richo]

It is just a tool. Teachers will not be replaced, they have numerous ways to actively assess a student for demonstration of knowledge. Simply use classrooms for any formative or summative assessment. Universities are overdue for assessment accountability. It is an innovation tool use it for progress not fear it.

[Alex-cwxf]

Aside from other implications of the uses of chatGPT—I remember Blue Book exams in college. No notes allowed, you have about 55 minutes to 1.5 hours to *hand write* one or more essays on a topic the professor picks at random. Even if that prompt was the same as a previous year's class, well, you still have to remember the arguments and bust that out on the spot. There were a few take home essays and exams but at that point everyone had the internet and was using wikipedia which everyone complained about, but anyone with a little bit of thinking get away with it and come up with plausible primary sources. I get that this won't work for all majors and is a major problem without in-person learning, but sometimes the way forward is to take a step backward.

[WeylandLabs]

Question : To ChatGPT - What's going to happen when bigger corporations just keep consuming smaller business entities forever ?

[VIDDTHEDESIGNER]

I wish this was a long-form conversation

[Overcaffenated]

- Open the pod bay doors, HAL.
- I am sorry Dave, I'm afraid I can't do that.

[danielmartin]

This unregulated industry needs to be checked. This is irresponsible, dangerous and alarming. Not only will tragedy unfold one day where we then hear "I thought I was talking to a human", but how does this actually teach people? It doesn't!

[hiratiomasterson]

There is definitely a LOT of tweaking that needs to be done with chatGPT - Prof. Galloway is a world leading expert in marketing, not AI, though the intersection of the two fields means it can be easy to blur lines of expertise. But this is early days, and the question we need to ask is "Where will this be in 2-3 years, let alone 5".

As for detecting AI created essays, this can be easily fooled by a human simply rewriting a fully written essay in some manner. All the hard analytical and descriptive work is done, it just becomes a bit of creative editing.

[IanPhilmore]

You can tell that guy is a damn good professor

[marbasfpv]

Exactly. Ive tried to use it for coding work, and it was constantly wrong. Picking and choosing function calls from unknown APIs when it was asked to use one specific. Interesting, but for productivity I went back to google.