In recent years, artificial intelligence research has accelerated, and people are increasingly interacting with AI. AI home assistants and self-driving cars Once upon a time, these were things from science fiction, but now they are becoming reality.
Some researchers and activists question whether AI is approaching the point of sentience, the ability to think and feel on the same level as humans. Some fear that sentient AI could overtake humanity, while others are concerned with subduing sentient life to do our bidding.
So how do we know if AI has become sentient? We will talk about the history of AI, where AI research is now, and how, and even if, we can determine if AI has crossed the line of sanity.
Key Findings
- Artificial intelligence research began in the mid-1950s with the search for artificial general intelligence.
- Today, most AI research focuses on specific tasks rather than general intelligence.
- Given our current understanding of consciousness, determining whether an AI is sentient may be nearly impossible.
A Brief History of Artificial Intelligence
Artificial intelligence has a long history, the first formulations of which were described in stories told millennia ago. Greek mythology speaks of Talos, a giant bronze statue that guarded the island of Crete and circled its shores three times a day. While the Greeks obviously wouldn’t have described Talos using the language we use to describe AI today, it’s amazing how long people thought about the boundary between human and machine.
However, it is only recently that AI has become something that humans can learn and develop. Many experts refer to 1956 as the official start of AI research. That same year, the Dartmouth Summer Research Project on Artificial Intelligence took place.
Over eight weeks, about twenty participants met and worked together to discuss and develop programs that could demonstrate learning abilities. The Dartmouth Summer Research Project is often cited as the starting point for modern developments in artificial intelligence, even if the eight-week event acted more like a brainstorming session. Programs created in later years taught checkers strategies, how to speak English, and how to solve word problems.
The US Department of Defense began actively funding AI research in the 1960s. Some researchers, such as Herbert A. Simon, have argued that within twenty years, AI will be able to do any job that a human can do. However, this prediction did not materialize, mainly due to the limited computer memory, and in the mid-1970s, funding was reduced. Research funding returned in the 1980s, but collapsed again in the second half of the decade.
The 1990s saw a second surge of research, this time focusing on more specialized and purposeful AI designed to solve specific problems. This made it easier for researchers to demonstrate success as their AI achieved tangible results in economics and statistics.
Increasing computer speeds, combined with the Internet and access to big data, enabled further advances in machine learning by the early 2010s. By 2015, Google was using AI in more than 2,700 projects.
Current landscape
Today, AI research looks very different than it did in its early years. Early research often focused on artificial general intelligence. People imagine this type of AI as humanoid, able to learn any task a human can perform. If you read or watch sci-fi media, this type of AI is usually mentioned.
Instead, many modern AI researchers are focusing on creating artificial intelligence to perform specific tasks. For example, deep learning is a form of machine learning that relies on large amounts of data and can mimic how humans acquire knowledge. People and companies can use deep learning for purposes such as speech or image recognition, recommendation systems, art creation, advertising, investment, fraud detection, and more.
AI research in general is now often treated as a separate topic from task-specific AI.
Current AI Products
If you’ve been around your TV in the past few months, you’ve probably heard of Open AI’s ChatGPT. This chatbot can take questions you ask and give you direct answers. This is a smarter approach to finding information on the web, as the chatbot gives you an answer right away, rather than a list of websites that can give you conflicting information.
OpenAI is not developed yet ChatGPT enough to replace journalists and people who write for a living. However, this technology has great potential and will fundamentally change many disparate areas.
You may have seen many people use another artificial intelligence product called Lensa last year. Users could upload photos to the Lensa app and—for a fee—get slightly tweaked animated photos of themselves to use as their Instagram or Twitter profile picture. While this is a rather futile use of AI, it shows just how ubiquitous it is becoming.
There are also many companies using AI for much more practical purposes. Retailers can use AI to learn where their supply chain is weak or demand is low and adjust accordingly. Insurance companies can use AI to identify escalation risk cases and offer potential solutions to avoid further conflict. The work of the customer support service may eventually be replaced by bots with artificial intelligence.
Launched some automated investment platforms using the power of AI to optimize investment for their users. With some apps, you can invest in a portfolio and have artificial intelligence move your investments to maximize returns and protect against downturns. This is especially useful because follow the news to decide what to invest in may take a long time.
The limits of intelligence testing and the Turing test
One of the big problems with knowing when artificial general intelligence became sentient is that intelligence testing is incredibly limited.
In 1948, the English mathematician, computer scientist and philosopher Alan Turing proposed the Turing test. This is a rudimentary method for determining the intelligence of an AI.
The test requires two humans and one AI. One person, the interviewer, has a conversation with two subjects, one person and one AI. If the interviewer cannot tell who is human and who is AI, i.e. the AI constantly tricks the interviewer into believing they are human, then the AI is intelligent.
Today, most experts agree that this test is ineffective for determining machine intelligence.
Another proposed method for assessing sensitivity is the General Assessment of Language Comprehension (GLUE). GLUE is similar to the SAT for AI, asking programs to answer English-language questions based on datasets of varying sizes.
However, even the GLUE benchmark and similar tests have limitations. Many argue that animals such as cats and dogs can think and feel, which is the basic condition for sentience. However, how many pet dogs can pass the multiple-choice test?
Also, with new developments such as ChatGPT demonstrating natural language processing (NLP) capabilities, it is becoming clear that some AI programs can process language. However, most people would agree that this is not the same as gaining intelligence.
How do we know if AI is intelligent?
Given the limitations of existing tests for determining intelligence, how do we ultimately know if a machine has acquired the ability to think and feel?
The truth is that this will be difficult and possibly impossible given our current understanding of consciousness. There is no consensus on the exact definition of whether an AI is conscious.
Research into tests that can prove sentience, as well as the science of consciousness itself, continues. Future advances may provide us with answers that we can use to better define and test sanity.
Will AI ever become sentient?
Another topic to consider is whether it is even possible for artificial intelligence to become sentient. Intelligent AI is a hot topic in science fiction, but could it ever become a reality?
Experts have taken an ambiguous position on this issue. Former Google engineer Blake Lemoine said that AI has already achieved intelligence through the Language Model for Dialogue Applications (LaMDA) chat program. In a conversation with the program, Lemoine stated that after reading the program, the program became sad. Outcasts and feared death.
Google said the claims were completely unfounded and fired Lemoine last year.
On the other hand, Associate Professor John Basle of the College of Social Sciences and Humanities at Northeastern University, who researches the ethics of new technologies, says: “Reactions like ‘We’ve created intelligent AI’ are grossly exaggerated.”
In an article for Northeast, Basle elaborates that he expects that if the AI ever gains sentience, it will have only minimal consciousness. He may be aware of what is going on and experience basic positive or negative feelings, similar to a dog who “does not prefer the world to be one over the other in any deep sense, but clearly prefers its biscuits to chewing” .
Researchers who believe in the possibility of intelligent AI are also debating whether this is a good idea. It is not hard to find people speculating about the various worst-case scenarios in which nefarious actors create millions or billions of bots to impose destructive political agendas on us. Anyone who has seen Matrix familiar with the media in which AI-enhanced machines turn against humans and eventually replace us as the dominant life form.
It remains to be seen whether this is nonsense or a potential future reality. Technology has evolved a lot over the last decade and it’s hard to say where it will be in the next.
bottom line
Artificial intelligence is an exciting field that people have been interested in in one form or another since ancient times. However, it has only become a field that people interact with on a daily basis in the last few decades.
Although there are many questions in this area, one cannot hide the fact that AI can perform many complex tasks. Many companies, from insurance companies to retailers, have begun to use AI to optimize their work.
Fast How do we know when artificial intelligence becomes sentient? appeared first on Due.