Why Would I Trust You?

You may have noticed a lot of hype at the moment around conversational services. Everyone is chasing the Star Trek computer. Apple has Siri, Facebook has integrated bots into its Messenger platform. Google has created a chat app just for the occasion of introducing bots. Amazon lets you talk to Alexa. Microsoft has built a platform you can integrate in any service. IBM has Watson.

Now that messaging is hot, and machine learning is evolving quickly, automated conversation seems to be a logical next step. Conversational bots can act as a flexible glue between people and complicated services that act on their behalf.

However, there are some big problems with it. In each of the following posts I will highlight an important issue and a possible solution.

Why Would I Trust You?

Did you know blurred pictures could be unblurred now using machine learning? So now that picture of you at the office party, slightly hammered, is not safe anymore.

There is no place left to hide. Even without using any social media, without having a blog or other website, you are on pictures with other people. You cannot live without other people and other services. And with machine learning, you can be found through the patterns of other people.

Considering how open our lives have become, we are becoming increasingly critical of who we give our information to. If you represent a company and company-sensitive data is being used, then you have an obligation to be extra critical.

The business model of most current online services is about competing on price in exchange for information. The truth is that they are not created for immediate sustainability like old brick-and-mortar business. Take any popular social network as an example. Your data becomes their currency. Your email address, likes, dislikes, search terms — all this data is of value to them. It is this data and network they aim for to give them a high future business and stock valuation.

I do not think the above model is sustainable. Our privacy or company information is tied into our personal security. As pattern matching through machine learning improves, we are increasingly becoming more exposed and insecure.

In essence, we have become used to giving our identity in exchange for convenience. We simply have to trust that our identity will not be misused and we are not compromised.

Trust is necessary for Caring Services

In the last post, I argued that to have good conversations, you need to know the user and their needs extremely well. This means harvesting a lot of information on them. Why would they trust you with that information?

Give Ownership and Control

It starts with the business model. If you do not pay for the service you get, or only pay a part of the cost, you do not own the result. This means that in the future, we will need to pay for services to access data we consider private.

On the flip side, by paying, we should get full ownership of our data and derived data. Ownership means we get the ability to access, change and remove information. In other words, we get to control our data. The service is fully transparent about what information has been gathered about you and what it does for you. You should also be able to export all this data in a meaningful form.

Have Transparent Communication

A transparent service should be able to be accountable to the user. The service needs to inform you why you got a notification or recommendation. At a higher level it means the service has to have open communication about everything it provides and contains.

Be Inherently Secure

Ownership also means that the information is only accessible to the owner. The information needs to be made extremely secure, not accessible to even themselves. As an example, take LastPass, a company that stores your passwords in the cloud. The user data that gets stored is encoded with the password that LastPass does not know. This means that even when they had a major hack, the hacker could not access any user passwords.

Trust Goes Beyond Technical

In the end, being trustworthy is a lot more than having stable and secure servers and saying you can be trusted (hello Google, hello Facebook!).

It is something that works from the inside out, from the business model and the company culture, to the technical implementation and into communication.

*Making your service trustworthy is essential for services that harvest your information for conversational services. In the future weekly posts I will describe other conversational issues and our approach to fixing them. *