[caption id="attachment_231332" align="aligncenter" width="1200"]
The introduction of ChatGPT has ramped up the evolution of generative A.I. programs, creating opportunities and challenges. | Photo by Sanket Mishra on
Unsplash[/caption]
I’ve been watching with interest as recent advances in artificial intelligence have reached fairly astounding stages.
For most of our lives, AI was something limited to science fiction depictions like HAL 9000 in “2001: A Space Odyssey” or Skynet in “The Terminator.”
[caption id="attachment_222223" align="alignright" width="300"]
Jacob Owens
Editor
Delaware Business Times[/caption]
All of that changed in November, when the California tech company OpenAI released its ChatGPT chatbot to the public. The program is what’s known as generative AI, or an AI that can create entirely original work ranging from question responses to essays, job applications to coding help, and much more.
It does so by relying upon its acquired knowledge after trawling the entire internet for data and developing speech patterns around how we communicate. Within a week of launching, 1 million people had registered for free to use the chatbot. In less than six weeks, that number topped 100 million, making it the first growing consumer application ever.
In comparison, it took TikTok about nine months to reach 100 million users and Instagram more than two years, according to data from Sensor Tower, an app analysis firm.
I began exploring ChatGPT a few weeks ago, and in many ways the appeal is obvious.
Have a strange assortment of vegetables left in the fridge but no idea what to make? Ask ChatGPT and it will respond with a suggested recipe outlining each step.
Don’t understand a specific topic and you want a quick primer so you can be knowledgeable in a conversation? Ask ChatGPT and it will give you a bullet point breakdown to get you up to snuff.
Short on time with Mother’s Day around the corner? Just plug in some basic info about the recipient and ChatGPT can cook you up a poem in a matter of seconds.
One of the most beneficial impacts that I think generative AI can have is the democratization of access to technology. While programmers used to have to get a college education in computer science – a system that inevitably underrepresents some minority communities and those below the poverty line – it may no longer be the case. With at-home training modules like Code Python and assistance from ChatGPT, an entrepreneur could conceivably create apps and programs on their own.
Joshua Wilson, an associate professor in the School of Education at the University of Delaware who has been studying generative AI, told me that it does feel like “we are at the cusp of a major change” in how technology will affect our lives.
He notes that small business owners could use generative AI to complete the more pedestrian tasks of their day like writing product descriptions for a website. That would allow them to spend more time doing more important tasks and potentially increasing their sales.
While Wilson said he’s primarily excited about the prospects of generative AI, he is also wary about its potential impacts too.
“Even though I'm like 60-40, that 40% is very concerning,” he said. “I do worry that the more that we think about these parts of our lives transactionally, and as sort of barriers to goals, that's troublesome because then it starts to see everything as sort of a commodity to just not be cared about.”
There are broad concerns about how learning is affected when tasks like writing and researching are automated. Generative AI isn’t also limited to text as well, as programs like Midjourney and Dall-E are producing original photos, and even more programs are beginning to branch out into video and music. How all of that affects our creative expression remains to be seen – why bother learning how to play guitar when an AI can cook up a fiery lick for your song?
All of these philosophical and ethical questions don’t even begin to breach the more troubling impacts that generative AI can have on intellectual property and misinformation.
One of the first examples that caught my eye was a rather innocuous photo that went viral online of Pope Francis wearing a puffy white winter coat allegedly made by the French fashion house Balenciaga. Except he wasn’t, and it turned out that the photo was created by an AI program, fooling millions like myself around the world.
Then another example popped up in recent weeks when an anonymous music producer used AI programs to write a new song and have hip-hop superstars Drake and The Weeknd sing it using voices generated by hours of voice samples. For a day, millions around the world believed the hit song of summer had arrived – only to find out the artists knew nothing about it and streaming services were pressured to remove it from the public.
While photos of puffy jackets or imaginative new songs aren’t terribly troubling, the opportunity for more damaging misinformation production via AI is apparent. It also could have impacts economically for companies and artists.
U.S. copyright and patent offices have ruled that work principally created by AI cannot be protected, but those with sufficient human authorship in conjunction with AI could, according to Joan Kluger, a patent attorney in Wilmington who has worked on AI products. That threshold for human authorship is not well defined though.
“They're looking into [generative AI]. It's going to take time for the framework to be established,” she told me, noting that the patent office hasn’t fully addressed how to deal with computer-assisted inventions, where generative AI would now fall. “It's a balance between what should be in the public domain for further development and what should be protected to benefit the investment of an inventor.”
These are concerns I expect to see debated for the foreseeable future. One thing seems certain though: we’ll likely look back at this time as a significant technology milestone.
