New Mission, New Website coming soon! Learn more now.

Equities logo
Close this search box.

Artificial Intelligence Needs To Get Smart — Part III

Are self-driving vehicles really a viable transportation solution?
Michael McTague, Ph.D., is executive vice president at private equity firm Able Global Partners in New York.
Michael McTague, Ph.D., is executive vice president at private equity firm Able Global Partners in New York.

Image source: Roberto Nickson / Unsplash

So far, this series has found that learning and agriculture — two critical issues for the future of the human race — are not turning up feasible opportunities for Artificial Intelligence (AI) to show its great power. This final piece of the series looks at another AI headline area — self-driving vehicles — and considers the problem AI faces from the nature of the high technology industry itself.
An assumption one often hears is that AI will someday control transportation. Imagine self-driving cars, sort of beyond mere electric vehicles. Railroads and airplanes controlled by the impersonal AI Robot, another of which will apparently serve peanuts on airplanes and punch tickets on trains by toddling down the middle aisle demanding payment, especially on their cousins: bullet trains.
Union Pacific ( UNP ) and Canadian Pacific ( CP ) must be waiting eagerly. Subways, ships, even bicycles and cars would move along flawlessly. No accidents, no late trains, no rude cab drivers, no problems of any kind. Clear messages about which track to find, where to board, infallible destination information, everything on time, etc.

Of course, in this digital dream world, many thousands would be out of work. In fact, early efforts to build such a system have proved disappointing. Crashes, human error and other factors make advances improbable because people always fear danger. As advocates of nuclear energy know, bad press damns. One report on self-driving vehicles shows 9.1 crashes per million miles traveled, compared to only 4.1 for conventional cars. 

Imagine a swift taxi running along a highway or a track while a slow futuristic bus runs along an orthogonal path. One electronic burn out may lead to a crash. The self-driving system would have few actual people on the scene: drivers (replaced by robots), police (replaced by digital cardboard manikins), medics (dislodged by electronic clipboards inside self-driving ambulances) to recognize the problem and respond.

As we saw with the discussion about AI and agriculture (Part 2 of the series), actual human successes involve a lot of hard work, expense and trial and error. The Los Angeles Dodgers baseball team was originally called the Trolley Dodgers in its Brooklyn days. That is because trolleys were prone to nearly collide with actual people. Robots can obey their programming and interact fairly well with other robots, but they have trouble dealing with the unpredictable reactions of actual people, much like the Brooklynites of old who ignored horns, whistles and the police and did things their own way. Human error would be everywhere and would generate considerable resistance to actually implementing such a dangerous system. 

The kind of world that would allow total transportation by AI would also require people to be docile. People would have to be reluctant to object if the self-driving vehicle killed a few dogs or even an occasional person. After all, such a system would be assumed to be error free. 

AI Not A Major Priority Among High Tech Giants

The prized high technology giants and how they operate form the ultimate barrier to AI development. Intel ( INTC ), Amazon ( AMZN ), ABB ( ABB ), Yaskawa Electric Corporation ( YASKY ) and IBM ( IBM ) got where they are by developing very attractive products and services: speedy computers, smart phones, superb software, rapid delivery, accurate payroll systems, robots, tracking devices, etc.

AI is something extra, not the center of what these rich companies do right now. The giants are looking to scrape up every small, promising AI company they can find. In part, they want to prevent another tech giant from owning AI and knocking them from their near trillion dollar valuations. This industry set up might lead to development of AI. It might also lead to using the acquired brains from the purchased companies in other areas, related to the core businesses of the current industry giants. Rather than rolling out a new learning system, the public might see robot, twenty-four hour, door-to-door, driven by truck, AI-managed deliveries. That would save a lot of money in the long run.

Years ago, companies resisted being acquired; now they generally welcome it. This set up may actually slow the development of useful AI products and services. The small companies with good ideas cannot resist being bought by a giant, cashing in and going into the future with attractive long-term contracts. The giants are not likely, however, to divert their focus on core businesses to create a new industry.

In addition, the current tech giants are already at risk of being broken up. Which one is a true monopoly? Microsoft ( MSFT ) has no rival in office software. Meta Platforms ( FB ) has bought many companies and has not been challenged by another tech giant as the social media champion, regardless of what it calls itself. (Keep on eye on TikTok, however, owned by Byte Dance). Meta has also been accused of colluding with Google ( GOOGL ) to avoid competition.

Investors realize that the tech giants are massive. Because people like posting on Meta's Facebook and searching on Google and because the Internet itself is free, few people worry about the massive size of the corporations that own these essential parts of day-to-day life. This could change, and the government, supported by a long legal tradition, stands ready to pounce.

Following this pattern, AI may also divide into separate camps. (We should say "silos" to use high-tech language.) Google will own the AI aspects in its search silo, and the others will have their own silos with minimal overlap. The government is also not an alternative for advancing AI. Washington is prone to waste money on unfeasible ideas. Private companies at least adhere to profit-making, which keeps them from wasting huge amounts of money.

We hope you found this series to be insightful, showing that the great promise of AI may be far more limited than its starry-eyed advocates believe. The next entry will move on to another issue in business and finance.

Michael McTague, Ph.D. is Executive Vice President at Able Global Partners in New York, a private equity firm.


Equities News Contributor: Michael McTague, Ph.D.

Source: Equities News