Winter is sleepy, and it is a time to prepare for a rebirth that will ultimately occur when spring rolls around. In regards to Artificial Intelligence research and advancement, an AI winter is a period where exploration into this field loses popularity but is quietly plugging along behind the scenes to re-emerge with new and exciting technologies.

Predominantly caused by a lack of widespread interest, during an AI winter, there is decreased funding into the field of Artificial Intelligence. With a lack of funds, research and development significantly slow down, and less innovation occurs.

Timeline of Artificial Intelligence

The term AI winter first got introduced during the American Association of Artificial Intelligence (AAAI) annual meeting in 1984. During the meeting, researchers Marvin Minsky and Roger Schank gave a presentation warning about the ebbs and flows the artificial intelligence industry has gone through since the field began to catch public attention in the 1950s thanks to Alan Turing.

While Turing is widely considered to be the father of AI, John McCarthy was the first person that coined the term artificial intelligence. In 1956, McCarthy organized the first Artificial Intelligence conference, which helped eventually lead to his founding of the first AI labs at Stanford and MIT. McCarthy developed many systems that are still used as the basis of AI and mainframe computer technology today.

The First AI Winter

The first AI winter began in the mid-1960’s after a disappointing letdown concerning language translation. During the Cold War, the U.S. government was trying to figure out the quickest and easiest way possible to translate Russian documents. The idea of a computer rapidly converting documents from Russian to English was an attractive thought, so research in this subject area quickly began.

After a few years passed without any real progress to show, the government and other investors began to raise concerns. Was there too much hype around what artificial intelligence was capable of doing?

In 1966, a report started circulating which stated that trying to use artificial intelligence to perform translations was much more costly than just employing humans to do the same work. After the release of the report, investors pulled funding, and research into machine translations came to a halt. At this point, the first AI winter began.

Difficulties in the ‘70s

In the early 1970s, AI research started to receive a decent amount of criticism in Europe. Reports highlighting the failures of Artificial Intelligence research began to surface, and criticism of this science spread rapidly throughout the world.

Initially, the Defense Advanced Research Projects Agency (DARPA) backed AI exploration with a large chunk of funding, but as criticism continued to build, DARPA officials began to realize that the original promises seemed overstated and unrealistic. Investors pulled a great deal of funding, and researchers who were studying AI were dealt yet another blow.

Good News in the ‘80s

In the 1980s, the first Lisp machines were introduced, and Artificial Intelligence became a hot topic once again. Lisp machines were, essentially, the first prototypes of home computers, and scientists and funders alike saw the potential of what could come about thanks to the invention of the Lisp machines.

Many governments throughout the globe showed interest in devoting time and money to further develop the vast world of Artificial Intelligence exploration, and excitement surrounding almost everything regarding AI began to skyrocket.

Since popularity was rising once again, the First National Conference of the American Association of Artificial Intelligence started in 1980 on Stanford’s campus. Four years later, Minsky and Schank spoke at the conference warning of the potential of an AI winter happening again—just like it did in the 70s.

Ultimately, the men were right, and another AI winter popped up as the Artificial Intelligence collapsed a few years later.

The late 1980s and early 1990s were a rough time for everyone invested in the AI industry. In 1981, the Ministry of International Trade and Industry in Japan put up a significant amount of money backing the research of an initiative known as the “Fifth Generation Computer Systems.”

The Fifth Generation computers were supposed to feature cutting-edge technologies and run far superior to any other machine that had come before it. Ultimately, after researchers had been working for about a decade, the project was a flop, and investors lost a lot of confidence in what the Artificial Intelligence industry was truly capable of doing.

Public perception of Artificial Intelligence was extremely low during the 1990s, and many people believed the industry to be well past its prime. Due too much hype surrounding what was truly attainable through Artificial Intelligence, another AI winter began.

It wasn’t until the early 2000s that Artificial Intelligence began to grow in popularity and credibility once again thanks to technology that was created to make machines seem much more realistic human-like. AI has been on a decent upward shift ever since thanks to inventions like robotic pets and voice recognition devices.

What Causes an AI Winter

Dips and peaks are typical in almost all markets. Even the railroad and real estate industries have had substantial upward and downward shifts. Artificial Intelligence innovation is no exception to the good times that come about during years of optimism, and the low points that occur when there seems to be a lull in discovery and progress.

Too much hype is a huge underlying factor of what causes the occurrence of an AI winter. Government agencies and investors tend to overestimate what can quickly come about from Artificial Intelligence research and the development of new technologies. When R&D is not cultivating new ideas rapidly enough, speculation occurs, and excitement for discoveries dwindles.

When the public opinion of AI research is low, investors choose to put their funds elsewhere by investing in other sorts of scientific advancements and technologies. A lack of funding causes new research to slow down, but it doesn’t mean that exploration gets completely halted.

Even during an AI winter, scientists are busy working behind the scenes to come up with new, useful innovations.

Issue of Public Perception

The exploration of Artificial Intelligence comes with some unique debates. Unlike most other technological research, AI often gets a reputation for being controversial. Many critics of Artificial Intelligence fear that one-day machines will become more advanced than humans, and we will lose control to robots.

While the fear of robots one day taking over the world may or may not a realistic worry, it’s a public relations issue that experts in the AI industry are forced to deal with to promote positive public perception.

If the public isn’t excited about a new type of research, there will be a lack of funding. Where there is a lack funding, there is less research and design.  Without advancements in R&D, the Artificial Intelligence industry will struggle to survive.

AI Winter in the Future

While AI winters may be natural to the lifecycle for the Artificial Intelligence industry, there are still factors that will help to keep lack of funding and public popularity to a minimum.

AI will continue to have success if it remains focused on providing cost-savings. Large companies, corporations, and governments are always looking for ways to save money, and AI innovation can bring about new ways to streamline production and to cut down manufacturing costs.

When large investors see the benefits of Artificial Intelligence, they will continue to funnel money into research and design initiatives by boosting up in-house research and providing grant-funded research opportunities to academic institutions.