Artificial intelligence (AI) and machine learning (ML) have become hot technologies in recent years thanks to their strengths in such areas as analyzing data, predicting trends, and powering virtual assistants and robots. IDC expects global spending on AI and ML to surge from $12 billion to $57.6 billion between 2017 and 2021.

Yet this isn't the first time companies fell in love with AI. Back in the 1970s, many companies and investors were enamored with AI-driven applications. But after several high-profile setbacks, many government agencies and venture capitalists pulled their funding, resulting in a so-called "AI winter" that lasted until the early 1990s.

A robot stands on an icy bridge with a little girl.

Image source: Getty Images.

Speaking to the Financial Times, Gary Marcus, a psychology professor at New York University, recently warned that "one of the biggest risks in the current overhyping of AI is another AI winter." Should tech companies heed that warning? Let's discuss some recent developments in the AI market to find out.

Crunching data versus "real" intelligence

IBM's (IBM 0.01%) Watson rose to fame as an AI-powered contestant on Jeopardy. The AI platform is now used across a wide range of industries -- including healthcare, cybersecurity, education, and even fashion. IBM often touts Watson' applications in hospitals, where it digests millions of medical records to help doctors make medical decisions.

An AI brain replacing a businessman's head.

Image source: Getty Images.

However, Watson is ultimately limited by the answers which have already been collected. It can't figure out unsolved medical mysteries on its own.

This "chicken and egg" problem caused many investors and medical institutions to abandon Watson, which they increasingly saw as a glorified indexing system.

Last year, MD Anderson Cancer Center cut its ties with IBM after spending over $60 million on a Watson project which was ultimately deemed "not ready for human investigational or clinical use." VC firm Social Capital CEO Chamath Palihapitiya also called Watson "a joke" on CNBC, claiming that IBM uses its "sales and marketing infrastructure to convince people who have asymmetrically less knowledge to pay for something."

The Facebook fiasco

Facebook (META -0.40%) uses AI to recognize faces, analyze a user's social profile to craft targeted ads, and more. But over the past few years, users, legislators, advertisers, and investors have become increasingly concerned about its data-mining practices.

In 2014, it conducted a bizarre emotion experiment on its users to gauge their responses to certain posts. In 2016, it unwittingly sold ads through its automated system to Russian agencies, which attempted to influence the presidential election with divisive posts and fake news stories.

Earlier this year, it was revealed that the data of about 50 million Facebook users was sold by a developer to data firm Cambridge Analytica, which was later hired by the Trump campaign. It was then revealed that Facebook was also tracking users' contact lists, telephone numbers, call lengths, and SMS messages.

Facebook now faces an FTC probe and a potential loss of advertisers and users over the fiasco. But this mess will also send shockwaves across the entire AI industry as the public starts scrutinizing the unchecked power of these AI-powered algorithms.

Really dumb chatbots

Facebook and Microsoft (MSFT -0.68%) both believed that AI-powered chatbots would eventually replace mobile apps or human customer service representatives. Those efforts didn't pan out.

Messages with a chatbot.

Image source: Getty Images.

Facebook's chatbots hit a failure rate of 70% last year, forcing the social network to scale back its efforts. Microsoft's experimental Tay chatbot, which it let loose on Twitter in 2016, started spouting out racist and misogynist replies within 24 hours. Microsoft quickly pulled the plug and issued an apology, noting that it had a made a "critical oversight" in its development.

Killer autonomous vehicles

Last year, the Trump Administration loosened regulations on driverless cars -- a main pillar of the AI market -- although critics claimed that the move could result in more accidents. Tesla, General Motors, Alphabet's Waymo, and Uber all saw their self-driving vehicles involved in accidents over the past few years.

However, Uber's self-driving vehicle was recently involved in the industry's first fatality after it failed to identify a pedestrian on the road in Arizona. That tragedy, which resulted in Uber being suspended from driverless tests in Arizona, will likely cause other companies to slow down their efforts. Regulators could also pass tougher rules on the industry.

That would be bad news for automakers or chipmakers, which are dependent on the rising adoption of driverless cars.

It's a cold front, not winter

These recent developments will likely chill the feverish enthusiasm for AI-related investments. But I don't think they will derail the entire industry, since AI technologies remain integral to many industries. Instead, an industrywide reality check is healthy, and could prevent another "AI winter" from actually occurring.