The world is very different now. For man holds in mortal hands the power to eradicate all forms of human poverty and all forms of human life.
John F. Kennedy
Humans have mastered many things that could change our lives, create civilizations, and ultimately destroy us all. I invented another one this year.
Artificial intelligence has been an impending technology for at least 50 years. A set of specific AI apps caught everyone’s attention last year as AI moved from the era of niche applications to providing innovative and useful tools. Doll E to create an image from a text prompt, Github copilot As a pair programming assistant, alpha fold Calculate protein shape Chat GPT 3.5 As an intelligent chatbot. These applications were seen as the beginning of what were most likely to be domain-specific tools. Most people (myself included) believed that the next version of these AI applications and tools would be incremental improvements.
we were very wrong.
This year, with the introduction of Chat GPT-4 You may have seen the invention of explosives, mass communications, computers, recombinant DNA/CRISPR, nuclear weapons, and other things with equal social impact all rolled into one application.if you’ve never played together Chat GPT-4stop and do so for a few minutes here. seriously.
At first glance, ChatGPT is a very smart conversationalist (and homework writer, Examinee).But this First software program have Become Can compete with humans on several common tasks(Looks at the link and realizes there is no turning back.) This level of performance was totally unexpected. even by its creator.
In addition to superior design performance, What surprised researchers about ChatGPT was that emergency actionIt’s a fancy term that means “we didn’t build it to do it, we don’t know how we know how to do it”. These are behaviors that were not present in previous smaller AI models, but are now appearing in larger models like GPT-4. (Researchers believe this tipping point is the result of complex interactions between the neural network architecture and the vast amount of training data it was exposed to. Basically, the year 2021 Everything that was on the internet as of September.)
(Another nasty potential of ChatGPT is its ability to manipulate people into untrue beliefs. ChatGPT “sounds very smart”, but sometimes it simply makes things up and does nothing even when the facts are not true. I’ve seen this effect on social media when it was humans who were manipulating beliefs. You can not.)
But that’s not all.
open pandora’s box
Until now, ChatGPT has been limited to chat boxes that users interact with.but Open AI (the company that developed ChatGPT) made ChatGPT Access and interact with other applications through APIs ( Application programming interface.) business side from an incredibly strong product application Even more incredibly powerful platform It can be plugged in and built by other software developers.
By exposing ChatGPT to a wider range of input and feedback, via APIdevelopers and users are almost guaranteed to discover new features and applications of the model that were not originally anticipated. An app that allows you to request more data and write code Doing it in itself is a bit chilling. This will almost certainly lead to new and unexpected emergencies. ) Some of these applications will create new industries and new jobs. Some make existing industries and jobs obsolete. And like the invention of fire, explosives, mass communication, computing, recombinant DNA/CRISPR, and nuclear weapons, the actual outcome is unknown.
Should I care? Should I worry?
First, you should definitely care.
Over the past 50 years, I have been fortunate enough to witness the creation of the first microprocessors, the first personal computers, and the first enterprise web applications. I have witnessed revolutions in telecoms, life sciences, social media, and more, creating new industries, markets, and customers literally overnight. With ChatGPT, you may see one more.
One of the problems with disruptive technology is that destruction doesn’t come with notes. History has it that journalists write about it but don’t recognize it (e.g. the NY Times published the invention of the transistor on page 46) or other journalists who don’t understand what they see ( For example: Xerox management ignored the invention of the modern personal computer with a graphical display) is overflowing. User his interface and networking at its own Palo Alto Research Center). Most people just stare at the giant d’s face.It looked like a toy to them, so they didn’t recognize it.
Others see the same technology and realize that the world will not be the same in that moment (e.g. xerox steve jobs). It may be a toy today, but they understand what will inevitably happen when that technology is extended and refined and there are tens of thousands of creative people building applications on top of it. increase. At that moment, they realized that the world had changed.
You may be seeing this here. Some immediately understand the importance of ChatGPT. others do not.
Perhaps we should take a deep breath and think about this?
Few people are concerned about ChatGPT or other ramifications. AGI-like I believe that applications and we are about to intersect Rubicon – Points of no return.They are 6 months moratorium proposed About training AI systems more powerful From ChatGPT-4. Others find the idea silly.
There is a long history of scientists worrying about what they unleash.Proposed by US scientists involved in the development of the atomic bomb civilian control of nuclear weapons. After World War II in 1946, the U.S. government seriously considered International Control of Nuclear Weapons DevelopmentAnd until recently, most countries non-proliferation treaty of nuclear weapons.
In 1974, molecular biologists discovered that a newly discovered gene-editing tool (recombinant DNA technology) Tumor-causing genes in E. coliThere was concern that in the absence of biohazard awareness and agreed biosafety best practices, there was a real danger of accidentally creating and releasing something with disastrous consequences. We asked for a voluntary pause in recombinant DNA experiments until we could agree on laboratory best practices. 1975, USA National Academy of Sciences sponsored what is known as Asiloma ConferenceHere, biologists develop guidelines for laboratory safety containment levels for different types of experiments, as well as lists of prohibited experiments (cloning of anything that may be harmful to humans, plants, or animals). Did.
Until recently These regulations prevent most biological laboratory accidents.
Nuclear weapons and genetic engineering favored unlimited testing and free control. “Let science take its course.” But even with these minimal controls, he kept the world safe from potential catastrophe for 75 years.
Goldman Sachs economists predict 300 million jobs could be affected By the latest wave of AI. Other economists recognize the ramifications of this technology. At the same time, new startups are forming, and venture capital is already pouring money into the space at a remarkable rate that will only accelerate the impact of his AI for this generation. Intellectual property lawyers are already debating who owns the data these AI models are built on. Governments and military organizations are beginning to grasp the impact this technology will have on the diplomatic, intelligence, military and economic sectors.
Now that the spirits are out of the bottle, it’s not unreasonable to ask AI researchers to spend six months following the model that other thoughtful and interested scientists have done in the past. (Stanford University Removed Version of ChatGPT Guidelines for the use of this technique should probably be developed in parallel with those for gene editing experiments. Risk assessment type of experiment and Biosafety containment level that match the risks.
Unlike the moratorium on nuclear weapons and genetic engineering, driven by the concerns of nonprofit research scientists, the continued expansion and funding of generative AI commercial enterprise and venture capital.
Welcome to our brave new world.
- be careful and persevere
- It’s a bumpy road
- is required Asiloma Conference for AI
- For-profit companies and VCs are interested in accelerating the pace.
Filed Under: Technology |