One of the challenges for Artificial Intelligence (AI) / Autonomous Solutions (AS) is to mitigate socioeconomic risks and negative impacts.
Financial Services have already experienced unintended risks when empowering algorithms to trade without sufficient checks and balances. During the afternoon of May 6, 2010, The Dow Jones Industrial Average (DJIA) plunged 800 points in less than 20 minutes.
This was caused by autonomous algorithms performing ‘program trading’ that led to the financial term Flash Crash. More recently, in October 2016, Sterling endured a temporary collapse against major currencies, which again was a Flash Crash, as autonomous trading
algorithms created their own market volatility.
Let's create a more universal definition of Flash Crash, which is applicable across all types of AI / AS:
“A Flash Crash is a very rapid, deep, and volatile deviation from accepted norms occurring within an extremely short time period. A Flash Crash stems from black-box algorithms, combined with high-frequency interactions or transactions, resulting with unacceptable
deviations or contaminated decisions.”
Other types of high-profile Flash Crashes have recently occurred:
- In March 2016, Microsoft’s withdrew a Chatbot in the USA called Tay only hours after its launch to engage with millennials when some unscrupulous Twitter users taught Tay how to be a racist. All credit to Microsoft, it had real-time controls in place and
withdrew Tay in a matter of hours from launch.
- In June 2016, a Tesla Model S with the Autopilot system activated was involved in a fatal crash. The car had not noticed a white trailer "against a brightly lit sky" so brakes were not applied. The algorithm "tunes out what looks like an overhead road sign
to avoid false braking events." Tesla reiterates that customers are required to agree that the Autopilot system is in a "public beta phase" before they can use it, and that the system was designed with the expectation that drivers keep their hands on the wheel
and that the driver is required to maintain control and responsibility for their vehicle.
- In August 2017, TenCent withdrew two Chatbots in China after they questioned the rule of the Communist Party and made unpatriotic comments. TenCent owns WeChat that has 938 million users.
Though Elon Musk and Mark Zuckerberg have not used the term Flash Crash their recent public squabble about AI / AS risks has been well documented. Musk at one stage described the Facebook CEO’s knowledge of AI / AS as
Elon Musk advocates that AI / AS needs to be proactively regulated to mitigate against a ‘fundamental risk to human civilization’. Musk uses AI / AS for transportation solutions such as for cars, space rockets and the hyperloop (proposed mode of passenger
and/or freight transportation). In these cases, the Musk approach is to ensure there are sufficient safeguards in place to protect people.
Zuckerberg, on the other hand says Musk is “pretty irresponsible” for using language such as
“I keep sounding the alarm bell, but until people see robots going down the street killing people, they don’t know how to react, because it seems so ethereal.”
The IEEE (Institute of Electrical and Electronics Engineer) like Musk are equally concerned, and have started the journey to develop a framework for designing AI / AS solutions. Their framework covers the following four areas:
· Principle 1: Human Benefit
· Principle 2: Responsibility
· Principle 3: Transparency
· Principle 4: Education and Awareness
The successes and benefits of AI / AS are already numerous. But like any significant advancement, each solution needs to be carefully designed, with proper checks and balances to mitigate against socioeconomic risks and negative impacts.