Investors must “dial back expectations” on self-driving vehicles as their mainstream adoption, unlike several other technologies, hinges on overcoming many complex and evolving obstacles, according to Morgan Stanley (MS).

In a research note, reported on by Barron’s, MS analysts Adam Jonas and Brian Nowak forecast that only 1% of all driven miles will realistically be traveled in a “fully autonomous” fashion by 2030. The analysts claimed that this predicted level, known as Level 5 by the industry, falls well short of what the stock market is currently demanding, adding that they frequently encounter investors “who are one to two decades more aggressive” in their expectations.

Jonas and Nowak said these investors generally fail to take into account the legal and ethical concerns surrounding autonomous driving technology, challenges highlighted by news earlier this week that a pedestrian was killed by a self-driving Uber car. Monday’s incident, the analysts added, should serve as reminder that autonomous vehicles (AV) are not like any other technology and cannot be compared to smartphone adoption.

“We believe a 1% Level 5 autonomous penetration of miles traveled by 2030 globally would be a very strong number,” Jonas and Nowak wrote in the note to investors. “We take this position to allow for the difficulty of the problem (that last 1% or last 0.001% of driving scenarios), the untested legal precedents, and the potential for regulatory and social forces to exert some influence on the process. Notwithstanding the urgency to address the loss of life and serious injury on our roads (>100 fatalities/day in the U.S. and over 3,500 fatalities/day globally according to the WHO), we believe comparisons between AV adoption and smartphone adoption rates are far from appropriate.”

Facing Complex Hurdles

Unlike other technologies, which can generally be brought to market as soon as they are fully functional, the analysts noted that AVs face much more complex hurdles. The legal and ethical challenges threatening to slow down adoption of AVs, they added, are almost impossible for the manufacturers of self-driving cars to get a handle on.

“We believe that you can't model morality,” they said. “We frequently ask investors and company management if they have a working assumption on the moral, ethical, legal, and regulatory frameworks surrounding fully autonomous vehicles. By far, the most common answer we get is along the lines of: “Not explicitly. We assume that it will work itself out over time.” We understand that it is not possible to model issues of ethics and morality. We only point out that this gap represents an important piece missing from the equation of market adoption.”

Jonas and Nowak also warned that AVs differ from other technological breakthroughs as it is more difficult to get consent from the general public to test them. (See also: Tesla's Auto 2.0 Near-Monopoly May Soon End: Morgan Stanley.)

“Individuals participating in a medical trial for an advanced cancer or diabetes drug are usually willing to accept the risk of serious side effects or death and have explicitly consented to participate in that trial; however, an unaware pedestrian walking down the sidewalk does not opt to participate in an autonomous trial in the same way,” they wrote. (See also: Waymo Leads Tesla In Self-Driving Tech Testing.)

According to Navigant Research, the leaders among the companies developing automated driving systems are General Motors (GM), Alphabet's (GOOG) Waymo, Daimler AG (DDAIF) and Bosch's partnership, Ford Motor Company (F) and Volkswagen Group (VLKAY). Tesla Motors (TSLA) was ranked last at 19th.