Open AI beta Launch

John J. Schaub

Dec 4, 2022 

I’m old enough to remember the very first time I used the Internet aka the world wide web. It was 1995 and I was talking to a high school guidance counselor about career options post graduation when he pulled me into his office to show me this amazing new technology that was going to open up all sorts of jobs. I was quite computer savvy and was familiar with BBSs (Bulletin Board Services - a precursor technology to the Internet for the kids reading this) so the concept wasn’t completely foreign to me but the ability to bounce between websites with a mouse click was amazing. My single sentence summary of the experience at 16 was ‘This is going to change everything’. I’ve been right about a few things in life but that one sentence response is probably the most right I’ve ever been. The Internet really did change everything.

My career in tech really got going as a Data Analyst working on Machine Learning based document classification and data extraction tools. Over the years I’ve led the development and deployment of a number of AI driven solutions mostly in the Identity and fraud prevention space. So as someone who is more than a little familiar with AI and ML I’ve been waiting for the day that the general public and frankly the tech industry itself arrives at its ‘This is going to change everything’ moment. I think it is fair to say that the public beta of OpenAI large language model is that moment. If you have not played with the beta already (and you really should) the text below is the response to the query ‘Write a blog post on the potential impact of AI?’

*******************

Artificial Intelligence (AI) is a rapidly evolving technology that has the potential to revolutionize many aspects of our lives. From healthcare to entertainment, AI is beginning to make its presence known in almost every industry. But with this potential comes a great deal of uncertainty, and the potential impacts of AI are far-reaching and difficult to predict. In this blog post, we’ll explore the potential impacts of AI, both positive and negative, and discuss how we can best prepare for the future.

The positive impacts of AI are numerous, and many industries are already beginning to take advantage of the technology. In healthcare, AI is being used to diagnose and treat illnesses more quickly and accurately than ever before. AI-powered robots are already being used in surgery, and the potential for further development in this area is immense. In the transportation industry, AI is being used to create self-driving cars, which could dramatically reduce the number of accidents on our roads and highways. AI is also being used in the entertainment industry to produce more realistic and immersive experiences for viewers, and in the finance industry to automate complex trading algorithms.

The potential negative impacts of AI are less well known, but no less important. As AI systems become increasingly powerful, there is a risk that those systems could be used for malicious purposes, such as creating automated cyber attacks or manipulating markets. Additionally, AI-powered robots could potentially replace human workers in certain industries, leading to job losses and economic instability. Finally, AI systems are only as good as the data they are trained on, and there is a risk that AI could perpetuate existing biases or make decisions that are difficult to explain or predict.

Given the potential impacts of AI, it is essential that we take steps to prepare for the future. We must ensure that AI systems are developed responsibly and are properly regulated to prevent misuse. We must also ensure that AI systems are trained on high-quality data sets to avoid perpetuating existing biases. Additionally, we must invest in education and training for workers in order to equip them with the skills necessary to adapt to a world where AI plays an increasingly prominent role.

In summary, AI has the potential to revolutionize many aspects of our lives and has already begun to make its presence felt in many industries. However, the potential impacts of AI, both positive and negative, are significant and must be taken into account when developing and deploying AI systems. By investing in education and training, regulating AI responsibly, and ensuring that AI systems are trained on high quality data sets, we can best prepare for the future and ensure that AI is used for the benefit of all.

*********

There are weaknesses in the text and I don’t see anything in the response that appears really novel or insightful but the fact is that the content generated is as good or better than the writing of many human beings and provides a solid summary of the topic. In simple terms if this was a question on an undergrad exam they would almost certainly get a decent grade. Many experts have made the claim that the way to avoid having your job made redundant by technology is to focus on creative work that cannot be automated but what we are seeing here is the first major efforts to automate aspects of the creative industry.

The obvious outcome here is that the already decimated market for any sort of low end written content creations (copy editing, technical writing, blogging ect.) is gone. My prediction is that essentially zero people will be making money in those fields within 36 months and those that do will be acting more as editors of machine created content then actual content creators. There is simply no way anyone will be earning money describing products or writing basic news reporting when an AI bot can mass produce similar content endlessly. There is a parallel and equally dire outcome on the way for low level graphic designers if you are following the recent results coming out of DALLE the OpenAI image creation solution which by the way generated the image you see on the left. 

And just in case you thought the answer was ‘To learn to code?’ I have some not great news for you. The query ‘Write a python script to take a series of comma separated values from a file called input.txt and print a list of all the values that are divisible by 7?’ generates the following response from OpenAI:

************************

#open the file called input.txt 

f = open("input.txt", "r") 

#read the contents of the file 

contents = f.read() 

#split the contents of the file into a list 

values = contents.split(',') 

#create an empty list to store the results 

result = [] 

#loop through the list of values 

for value in values: 

#check if the value is divisible by 7 

if int(value) % 7 == 0: 

#if it is, add it to the results 

list result.append(value) 

#print the list of values that are divisible by 7 

print(result)

***********

Again you can nitpick about the approach but it does work and it is solidly better than many undergrad responses to a similar question (it is commented!) and there is no question that tools like this will continue to improve as the economic incentive to replace or augment a software developer is vastly higher than a copywriter. If you want to have your mind blown make a deliberate error in the code and ask OpenAI 'Why doesn't this code work?'. I do not code as the main part of my job but in my life and career I do so occasionally and I can promise this tool will change my workflow completely.

So where do we go from here? In my mind the impact is going to be far far bigger than just the low level text and image creation jobs or entry level software development. There will come a time in the not distant future where content creation will move from text and still images to video. And the day I can enter the query ‘Create a new Indiana Jones movie starring Harrison Ford?’ and get a response a $200B/year (in the US) industry will evaporate. And to put it in perspective that day might come in as soon as the next five years. People can make the counter argument that there will always be a market for human created content, human created paintings and human acted films and I agree there will be but there is still a market for hand made cars that doesn’t change the fact that the vast majority of vehicles are built on massive assembly lines with significant automation using a fraction of the human labor that hand building would require. Many creative industries including technology are going to go through the same mechanization process that heavy industry went through a century ago over the next decade or two. The first areas to collapse will be the low end, the mass produced articles, the clipart style images, the hallmark movies but in the end even the major blockbuster films and award winning novels will be under threat as well. In short ‘This will change everything’.

My advice to anyone, particularly anyone starting their career, is to get very familiar with the emerging tech and figure out how you can use it to make yourself better at your job. At the very least the added familiarity will give you a 24-48 month advance notice of the truly earth shattering changes headed to your industry and you as a skilled user will be among the last ones made redundant. For my part in my free time I teach in a post graduate business program and rather than try to stop students from using these sorts of solutions to complete assignments I’m going to require it. The reality is we will not be able to avoid the impacts so being aware that they are coming and skilled in there application is the best we can do.

There are some massive questions that I have not even touched on in this post, the biggest being the legal aspect, specifically intellectual property. If you understand even the basics of AI you know that these systems have been trained on a massive corpus of knowledge which is in effect the sum total of human culture collected over thousands of years. Will massive companies be allowed to exploit that collected knowledge for free? Will there be a tax implemented on AI? Honestly there is no clean answer. Historically we have not taxed creators who leaned heavily on previously created works so I'm not sure that there is any precedent to do so here. That said, the legal system absolutely bars the use of copyrighted and trademarked content and it doesn't appear that OpenAI has done anything to detect when its output might infringe existing legal protections. Governments are going to be pressed pretty hard to at least slow down the freight train that is coming because I cannot imagine a company like Disney would be comfortable with an AI happily answering the query 'Write a script for an Indiana Jones film called Indiana Jones and the Dial of Destiny?' as OpenAI will.