Mike Russell

OpenAI just disbanded their risk team (see link below).

It’s hard to tell what the effects of OpenAI’s organizational change will be, given we can’t see everything going on inside the company.

However, it’s not hard to tell what companies adopting/adapting AI and AI tools need to do …

More needs to be considered beyond just understanding AI and figuring out what to do with it – the direct opportunities and implementation.

Taking advantage of any technology advance is never about just the technology.

It’s also about the implications and downsides, not just the potential upside.

Risk evaluation and mitigation isn’t just for OpenAI – it’s for everyone.

If you’re currently using AI or thinking about doing so, consider:

• Ethics … how does AI support or erode the company values and culture? Of customers and other stakeholders?
• Legal factors … what are current and proposed laws regarding AI specifically?
• Compliance … if there are other regulatory factors involved (e.g. as in banking, medical fields), how will AI help or hinder compliance?
• Bias … how will we ensure how we implement AI does not create or induce biases that are contrary to our intent and culture? Contrary to law and complaince?
• Other risks … what are the effects to other factors like security, safety, misuse (by customers AND internally), etc.

Blinded adoption is like blinded driving … your chances of a wreck are usually higher than you’d like.

 

https://www.linkedin.com/news/story/risk-team-disbanded-at-openai-6741066/

A recent post on LinkedIn (https://lnkd.in/ecrcJRWM) commented on how job postings seem to leave out critical aspects of a leader’s success:

While a number of organizations chase after leaders with the right position title or number of years sitting in a particular role, very few seem to appropriately acknowledge and reward leaders with a proven track record of rehabilitating, building, and nurturing healthy, high-performing teams. Not only do the engineers/individual contributors on the team appreciate those skills, but the decrease in attrition, increased technical performance, and accelerated growth of the team as a result of these highly-effective leaders directly contributes to critical business outcomes.
So, why aren’t these leaders recognized?

Some initial thoughts:

• This will likely get worse with indiscriminate AI searches and screening for specific titles, industries, and other demographics … worsening tunnel vision in recruiting.
• Focusing on demographics is easier than looking at substance, even before AI.
• While many trumpet “culture eats strategy for breakfast” or similar memes, not many companies hire that way nor do leaders lead that way.
• Internal turnarounds and team building can be huge endeavors, but since the starting point is often in the negative zone/hole, the huge results may not look as big compared to moving from a positive position to a higher positive one.

Questions to consider:

❓ Will AI make finding the right leaders better? Worse?

❓With the huge number of applicants and resumes floating around, how can HR folks and recruiters efficiently find the gold needle in the dross haystack?

❓ What’s the best way to express/make visible the value of rehabilitating, building, and nurturing high performing teams?

Communication problems can frustrate your customers … and your employees … and eventually your results.

Getting business “right” depends on a key foundation: communication.

That is a business problem – not a person’s problem – at root.

Effective, efficient communication can’t happen when there are differences in language and understanding.

One would think in the age of automatic translation capabilities, emphasis on assimilating new hires into organizational culture along with training and development, etc. that communication issues would be mostly gone.

Not so.

Here’s an example:

We recently had a credit card stolen and notified the issuer – one of the BIG ones.

We went online and flagged fraudulent transactions that began showing up immediately. We then had two calls with the card fraud department to sort out the valid charges from fraudulent ones.

Then, more than a week later, I got a call from the fraud department wanting more information.

I had different language options (good for customers) and chose English for the call.

The resulting conversation was, in a word, difficult … I should also add irritating.

The card representative could not communicate well enough in English to get the call done, much less well.

A call that should have been 5 or less minutes in duration, took about 25. I still wasn’t sure everything is correct.

This was NOT a specific person problem from a root cause view.

This WAS a company problem.

A problem someone not matching the person to the task. A problem with hiring and selection … with training and development … with performance evaluation … and the rest of the “people cycle.”

There are at least 3 takeaways for your business illustrated by this situation:

First, make sure the language skills are adequate for the task.

Different tasks require different levels of language comprehension.

On this call, the representative couldn’t handle a basic conversation in the selected language.

For instance, even dates were hard to get right. The representative could not understand dates in a format like “March fifth.” I switched to saying with a number like “March five” … and that only sort of worked. Took multiple times to get ONE date right. And there were many to go through.

Add in the representative had a very strong accent difficult for me to understand … as probably mine was to her.

It would have been comical if it wasn’t about lots of money.

Same thing happened with numbers like charge amounts.

I resorted to “spelling” out the numbers like “one two three” rather than “one hundred and twenty three.” Decimal points were almost impossible to get across. Changing to using the terms “dollars” and “cents” didn’t help much.
It took much trial and error with lots of different phrasing to get basic facts sort of straight.

This representative’s language skill was inadequate for even basic bank teller transactions … much less fraud investigations.

Second, make sure thinking skills are adequate for the task.

Thinking skills underlie communication skills.

In the case example, I was working with one of the big credit card issuers about a stolen card.

One of representatives couldn’t converse in the language offered by the business, so even dates and numbers were a challenge.

It got worse.

The representative apparently was focused on one task … find a date after which all transactions were fraudulent.

There were some major problems with this, common to any credit card loss or fraud:

1. Some of the charges on the old card were valid on/after the fraudulent transactions. These charges were being lumped into the fraud category only by virtue of being after the actual fraudulent charges.
2. New cards had been issued, and there are new valid charges against the new cards. These charges were also being lumped into the fraud category only by virtue of being after the actual fraudulent charges.

Based on those problems, there was no way to set an arbitrary date after which all transactions were fraudulent.

I tried multiple times and multiple ways to describe and illustrate the problems … and why just “a date” wouldn’t work.

Once we got past the basic language comprehension barriers, there was still a logic/thinking problem. The representative did not understand the issue and stuck to the “single date” task.

Only by being insistent was I was finally able to get agreement to go through charges and flag valid/fraud one by one.

The level of understanding and thinking required to deal with this situation was higher than just getting a “date” for the file.

There are several cause possibilities:

• The business provided a script that failed to account for the common situation above.
• The business did not allow deviations from the script to accommodate situations outside the basic script.
• The business did not screen for ability to go “off script” even if allowed.

It is wishful thinking that all situations can be covered by customer service scripts. Customer service (and other roles) need appropriate language and thinking skills to handle the “off script” situations successfully.

Third … and deepest/biggest on the “iceberg” of effects … there are broader aspects to address and work in a cross-cultural environment.

Things like vision, employee growth, and trust are fundamental to employee engagement. And customer satisfaction in turn.

And all depend on effective communication.

There is more than mere “word translation” at play when there are language gaps.

Translating words into another language is a starting foundation. Understanding the meaning of the word is next. Understanding words together – along with any idioms – is the next level above that. Fitting all that “translation” into a particular business context is even higher up.

These kinds of communication issues are also culture issues. They are not just due to someone being from another country. Different subcultures within a country can create similar issues.

Different educational backgrounds, different life experiences … many things can create different “languages” and understandings that are a barrier to communication.

And that’s all before we even consider the ladder of inference 😉

Don’t miss communication disconnects when trying to improve most anything in an organization.

There have been several articles in recent months about AI replacing executive jobs.

Here’s one where nearly half of CEOs believe AI could replace their jobs! https://lnkd.in/dt2C7JHh

I keep coming back to: huh?

If that’s the case, why are they in the jobs right now?

Is what they do really fully automatable? Or is this just going along with the “AI flow,” an attempt at organizational egality PR, or something else?

Also from the linked article: “During Agarwal’s tenure as CEO of edX, he spent nearly 80% of his time on “mundane” tasks like “reports and repeated presentations, or saying the same thing to a lot of people in different ways,” he says.”

First, communicating mission, values, and vision is a big part of a CEO’s job, along with ceaseless working to keep all stakeholders aligned. Yes, that may seem like “grunt work” but it’s essential to business success. Did he identify what was his unique “add” and delegate the rest, to human or AI? Doesn’t sound like it.

Second, if the tasks were truly low value, that’s a MASSIVE misuse of CEO time, unless the CEO is paid the same as an executive assistant or other roles – whether all human or human/AI combination – that could be doing “mundane” tasks where “mundane” means not central to the CEO’s value to the organization. Unless of course by “mundane” he means “I don’t like,” and it’s actually part of the role 😉 … back to the first point.

A CEO needs to focus on what a CEO should focus on …

A more likely scenario is a lot of small(er) companies with human CEOs and substantial work being carried out by AI.