This article is the second of a three-part irregular series on the regulation of robo-advice and the liability it poses to wealth managers. Read the first article in this series here.

Robo-advisor platforms may well be attractive to advisors seeking new ways to cut costs, streamline their asset allocation models or engage with younger clients. But some industry lawyers warn of a burgeoning tide of liability for advisors who incorrectly use these dispassionate tools.

The legal state of robos is still somewhat unclear. All the regulation which applies to human advice also applies to advice provided by a computer, but how that would apply in practice might not be so simple to ascertain.

In general, liability for an advisor only arises if the investor incurs a loss. Investors can’t simply sue an advisor because they don’t like the advice or service they receive, whether it be from a person or a machine. Jack Murphy, a partner in Dechert’s Washington, D.C. office, says there are only very limited private rights of action against advisors under the law governing financial advice – the Investment Advisers Act of 1940. Murphy says it comes down to state contract law if the investor feels there’s been a breach of the advisory contract.

He says liability depends on the nature of the contract breach and whether the actions of the advisor – or robo-advisor – violate the standard of care in the contract, which may be simple negligence or may require gross negligence, depending on the jurisdiction.

“There doesn’t seem to be any difference between leaving a robo unsupervised and a failure to supervise a subadvisor,” which the SEC has held to violate section 206 of the Advisors Act, says Murphy. “An advisor can’t just hand a portfolio over to a subadvisor or robo and wash its hands of it.”

Attorneys specializing in wealth management counsel FAs to fully understand the mechanics of the robos they are using – just as well as they should a subadvisor – whether the robo be a system they’ve built themselves or one they’ve licensed or purchased from a third-party vendor.

Wayne Zell, an attorney at Reston, Va.-based law firm Odin, Feldman & Pittleman, questions whether using a robo to develop an investment plan for clients is really any different from “using your own mind.”

“The advice you provide a client is only as good as your own skills, experience and methodology,” says Zell. “When it comes down to it, there’s not a lot of difference between a bad robo and a human advisor who isn’t good at their job.”

One industry watchdog has recently sought to clarify its position on how it thinks FAs should use automated investment advice in their practices.

While emphatically shying from creating new legal requirements for broker-dealers, Finra last month attempted to outline its position on the technology in a report which laid out best practices and underlying principles for using robos.

It’s key message? Any robo a broker-dealer uses will be held to the same suitability rule to which human B-Ds are held when they provide investment advice. A robo is not a substitute for knowledge.

Finra says it’s up to the B-D firm itself to make sure they understand the key assumptions – and especially the limitations – of the robos they choose to use. The onus is on the broker-dealer to determine if a robo isn’t appropriate for a particular client.

But Douglas Rappaport, a partner in New York-based Akin Gump’s litigation practice, says it’s not enough for an advisor to only conduct due diligence when it’s buying a robo: “You can’t just leave it alone. It requires monitoring, and must be checked and retested in normal market conditions and under stress, especially if the robo’s job is not only to recommend asset allocation but also rebalance portfolios. If an algorithm isn’t choosing properly for technical reasons, then you will have a misallocation in the client’s portfolio.”

At the core of every robo-advisor is an algorithm – a mathematical formula – which calculates asset allocation based on inputs like age, income, goals and risk profile, says Zell.

For advisors using robos, Zell thinks the primary liability would likely be if the underlying algorithm malfunctions; if it doesn’t do what it’s supposed to and subsequently loses an investor’s money: “An algorithm is only as good as its programming and it can’t anticipate everything. It doesn’t matter how smart the people are behind it.”

Rappaport likens the potential liability exposure of using robos to that of errant algorithms in the electronic trading world, where bad or malicious programming has been known to crash markets, such as the so-called Flash Crashes of May 6, 2010 on the New York Stock Exchange and on the Singapore Exchange in October 2013.

“When it comes to the use of algorithms anywhere in the financial system, the abiding concern of regulators is a ghost in the machine. When you withdraw the human element it makes the process non-sentient,” says Rappaport. “We live in a world where computerization is seen as perfect and pristine, but that’s not always the case. While less emotion in investing may sound attractive and automation is certainly cheaper, the potential exposure an RIA leaves itself open to can increase significantly without the proper monitoring.”

How robos handle risk – and the advisor’s understanding of that handling – is an important area for Finra. It warns firms that they are responsible for how the robos they use assess suitability for a particular investor profile. Firms must be able to understand how the robo appropriately calculates the portfolio’s risk, return and diversification, all the while avoiding or disclosing conflicts.

Wayne Zell

Zell recommends checking for potential conflicts of interest between the advice pushed out of the robo and the underlying investment vehicles offered.

“There may be conflicts within the robo advisor’s sponsor or owner which the advisor doesn’t know about,” he says. “Particularly with third-party robos, the advisor might not be aware of what the fees really are.”

And not only must broker-dealers ensure their robo fully assesses the risk capacity and risk willingness of investors, the firm must also be able to understand and address any contradictions and inconsistencies in a client’s answers.

“If you’re purporting to provide investors with a customized portfolio, then the SEC would likely say that asking one or two questions is not enough to build an adequate risk profile of the client,” says Murphy.

“It appears that some of these robos may be collecting fairly limited information on their clients,” says Jason Ewasko, compliance director of Cipperman Compliance Services. “Whether it’s adequate doesn’t just depend on the design of the questionnaire – although this is extremely important – but also on how the investor decides to answer. The more automated the process, the more potent the potential liability that the advisor doesn’t truly understand the client.”

Finra tells broker-dealers to make sure their disclosures are up to snuff and deal specifically with how their robos work, including rebalancing procedures and reactions to major – or unforeseen – anomalies in the market.

And ultimately, investment-specialist attorneys see fulsome disclosure as the best way to mitigate most risk which robos pose. But it isn’t a silver bullet.

“Robos are held to all the same standards as traditional advisors,” says Ewasko, warning that pretty much anything that can go wrong with human advice can go so much more wrong with robo-advice.