Are we at the tipping point of an AI winter or has it channeled over the years?
Rethink Robotics, a Boston-based robotics company founded by Rodney Brooks in 2008 had to shut down the company due to poor sales and acquisition. This news was a disappointment for the artificial intelligence industry.
Rethink, a manufacturer of Baxter and Sawyer (two robots) was one of the leading voices in the robotics and AI domain – a bridge built to usher in the new era of cooperation taking place between humans and robots. Having to hear the company shut down was dismay for all the other AI industries.
“Scott Eckert, CEO of Rethink Robotics said, the company’s closure is not indicative of wider troubles in the industry. But the shutdown of the company is not an isolated event. In August, Mayfield Robotics, the manufacturer of the Kuri home robot, shut down after failing to find a workable business model.”
In September 2018, The New York Times published news that disclosed a few of the inside details about Boston Robotics, a company that created hype by uploading a lot of videos about robots. While this company was open about publicizing the achievement of their robots, but what The New York Times projected was whether the robots meant business, or was it displaying a research lab?
As a result, it was still not clear what type of real-world problems these robots can solve.
In other domains of AI, we’ve all come across news where companies create hype about how AI can deliver different tasks but have failed to disclose their capabilities.
Ever since the AI inception, the industry has undergone many hype cycles. In the last decade, underdelivering and overpromising of AI leading to what is called the “AI winter,” a phase where lack of interest was shown toward the research in AI, which eventually led to lack of funding in the field.
The First AI Winter:
The year 1970 was when the first AI winter took place. During this phase, leading researchers and scientists made promises about the efficiency of AI, and how it would outmatch the thinking of humans in chess, checkers and automate translations of written texts in different languages – however, none of these promises materialized during that period, it was only until a decade later.
The Second AI Winter:
While the effects of the First AI winter started to decline, a new era of AI started.
The 1980s – 1990s.
This time, it wasn’t about the promises of AI being efficient and outmatching the way humans think, but, this period was all about creating commercial products. And at the heart of commercialization was the thought about AI systems and how the technology can help financial planning, geological exploration, microelectronic circuit design, and medical diagnosis.
Now, these systems worked by creating a rule called the “if-then rule” to set accordingly and by surveying experts. This approach was called the “top-down” approach to AI – an approach that believed to be the best way to create AI.
- In 1984, a magazine called Business Week joined the AI hype and published a headline that reads, “AI: It’s Here.”
- Similarly, other companies made claims that said, “We’ve built a better brain” stating “It is now possible to program human knowledge and experience into a computer … Artificial intelligence has finally come of age.”
Are we at the brink of another AI winter?
While skeptics persist, many people believe the AI revolution is underway.
Katja Hofmann, a principal researcher at Microsoft Research in Cambridge says, “I have the sense that AI is transitioning into a new phase.”
It would be unlikely to say, we’re getting into another phase of AI winter since billions have been invested in AI, and also considering facts about recent breakthroughs in AI. Some researchers say it would be wrong to call this phase – an AI winter.
Noel Sharkey, a professor of AI and robotics at Sheffield University and a Robot Wars judge told the BBC that he likes the term AI autumn – while several other researchers agree.
How will AI look like by the end of the 20s?
According to Catherine Breslin, an ex-Amazon AI researcher, she says, the next decade will see a more realistic and measured view of AI’s capability as compared to the hype seen in the past decades.
AI is evolving; thus, it is difficult to come to conclusions this early.