Silence of the lambs

I’ve been silent on this blog recently. At least compared to my historical pace of publishing posts.

One explanation is that I’ve started writing my Perspectives on Power Platform newsletter. If you’ve been follwing my blog for the technology topics around MS BizApps, I recommend you to check out perspectives.plus and subscribe to receive the newsletter issues via email. (Why use a newsletter instead of a blog? That’s a subject for a future post to come.)

Silencing your employees

There are other types of silence. The one that got me reflecting on these thoughts is what has been written about the NDAs at OpenAI. In short, the organization has imposed very strict contractual terms on departing employees. The exceptional issue appears to be OpenAI claiming the right to claw back vested equity. This right would be triggered if the ex-employee would criticize their former employer – ever. With no end date.

Reading about these reported policies has caused me actual physical discomfort. There is just something about the pre-emptive silencing of the people who work at an organization that rubs me the wrong way, in a big way. Can there be a more obvious way to state “we won’t really ever trust you” than imposing something like this?

In this case, the executives have of course played the get-out-of-jail card of unawareness. From the Xeet of Sam Altman:

Many will surely believe the explanation. That it has simply been some lawyers out there who’ve been zealously protecting and pursuing the client’s legitimate interests, within the bounds of the law. The corporate way: we’re doing all this just because it’s the way corporations work.

Weapons of mass distrust

The fact that OpenAI has never clawed back money from ex-employees is completely irrelevant. The impact of such contract clauses takes place regardless. The whole purpose of nondisclosure agreements is to stop something from happening. They are the corporation’s nuclear missiles.

While there are perfectly legitimate scenarios in various business relationships where NDAs enable confidential discussions to take place (everywhere in consulting, for example), this one is very different. When you have a policy that forbids criticizing the company for the duration of the former employee’s lifetime (and also forbids them from acknowledging the existence of the NDA), this is not about establishing trust. It is a weapon against another party that you by default do not trust. Period.

This model establishes a system of silence. Both before and after an employee leaves the organization. This is because it’s important to understand how the concept of criticism is defined. It is not something that the employee (subject) can evaluate. It is unilaterally defined by the corporation (object).

From this situation arises the imbalance of power that can impact organizations in everything they do. If your employees must be continuously evaluating in their heads the question “could someone interpret what I’ve said as criticism”, they will only say out loud a small subset of what they really think. Self-censorship is a destructive pattern that can repress any initiatives for building trust among teams.

As the information worker organizations increasingly become independent from physical locations, our communications start to become mostly digital. No matter if its emails, chat messages, online meetings – our modern multimodal AI algorithms will convert everything into text. Potentially storing it forever. Making it available for queries, in a whole different context than where the communication initially took place.

In such an environment, where do you create room for the informal, uncensored discussions to take place? This is a very hard problem to solve in practice. That’s because the root cause isn’t the traceability of digital communications. The need for creating a separate space where people can express their thoughts and feelings is the problem. Such separation should not be needed to begin with.

Choosing transparency

Ever since the Web 2.0 era tools and techniques became available, I’ve been a vocal proponent of working out loud. The idea that you should be proactively making your work visible to the networks through which value can eventually be created. Not just reactively providing specific information when requested. Making everything you type as broadly visible as possible in the given context.

Why bother? Because we ultimately should be conscious of not wasting the keystrokes we have left in us:

For someone like me who believes in the transformative power of radical transparency, any organizational barriers that encourage silence are a problem. Most of them are softer barriers, such as the general convention of how people around you behave. Others are technical barriers that results from the design of our information systems – be it intentional restrictions or unintended practical limitations. Finally, there are the contractual weapons mentioned earlier.

I don’t believe there is a way to separate external transparency and internal transparency when it comes to company culture. By external I’m referring to communication that takes place out there in open communities and networks that connect professionals from several different organizations. The internal part is about all communication that takes place within the (fire)walls of an organization – voluntarily, without an explicit process to require such activities to take place.

If internal transparency is not something that is organically allowed and encouraged to grow in the organization, you’ll likely have to try and force the external transparency. Meaning, it’s hard to get your experts to actively participate in community activities and share their knowledge with the outside world if it’s not a pattern that exists internally in the corporation. There will always be exceptional individuals, but it will not become a part of your culture. Thus you cannot leverage the network effects but rather have to pay to get people to notice your company.

The excuses for silence

Transparency can be a virtuous cycle for the business. Silence is often a vicious cycle. Why isn’t every organization then gravitating towards a more open and trusted culture of communication? Boy, that’s a big question that requires some serious investigation. Or maybe just throwing a quick question at ChatGPT – which provides the following reasons:

• Fear of negative exposure
• Control and power dynamics
• Short-term focus
• Lack of trust in employees
• Cultural and structural barriers
• Legal and regulatory concerns
• Inertia and status quo
• Risk aversion

Going through the list and the more detailed explanations under each item (try it with your AI tool of choice), it’s easy to see why choosing transparency isn’t by any means easy for organizations. Individuals exist also in the leadership team, thus the choices made in managing a company aren’t made simply based on cold, hard logic. We all need to feel psychologically safe at work, among our colleagues as well as with our managers/subordinates. If we’re emotionally or physically drained, that sense of safety is really difficult to reach. And so the cycle begins.

Transparency rarely just happens, yet silence is easy to achieve. Like in the example from OpenAI. The fact that the word “open” is included in the company name has been the source of ridicule for many reasons (closed source models, lack of respect for copyrights). Even though they’ve got the technical and financial resources in place to achieve top results in the global AI race (largely thanks to Microsoft), the company’s culture of silence can turn out to be a significant handicap in the long run. It’s certainly not a place for everyone to feel safe at.

How you treat the employees who are moving on is a signal of whether you consider them to be a potential future asset or a liability. As the detailed report from Vox.com reveals, there’s not question what side OpenAI’s culture falls on:

“Just doing things by the book.” Just assuming that whatever the employees do in their professional lives from here on could not possibly be of value to the organization. As opposed to the clear and present risk of them talking with others and expressing their own thoughts. Talk about short-term focus in a networked world.