Responsible, by ClearOPS

Did you miss me? I was skiing in Jackson Hole last week and visiting friends, so I let it go - this newsletter that is. Now, I am back and I have a lot to say!

Being responsible is not easy, but without a dialogue, it is impossible.

What I have for you this week:

• Snippets - Thoughts on an AI conference

• Caroline’s weekly AI Governance tips

• Chef Maggie Recommends

• AI Bites

Unlock the full potential of your workday with cutting-edge AI strategies and actionable insights, empowering you to achieve unparalleled excellence in the future of work. Download the free guide today!

Yesterday and today, I attended an AI conference called Human[X] and it confirmed my theory. “AI Agents” is a made up term (whether by marketing or VCs, I am not sure), but the interest in them is significant. Every AI agent session at this conference was packed! Particularly those focused on AI agents replacing humans in the workforce. However, I have a feeling that it isn’t because a bunch of CEOs were looking to use AI agents. I actually think it was people genuinely concerned that these agents were coming for their own jobs. Self interest will always trump. Which, if you are talking about an agent, is exactly my point! An agent is supposed to make decisions in the best interests of its principle, not itself. That is why these things need regulation. I will bring more thoughts on this to my Regulating AI Agents class. Interested? Take this quick survey to see if you are a good candidate.

I decided to go to this conference because I was hoping to see the latest AI tools. While I did see some interesting tools, most of the AI is still at the ChatGPT level of problem solving, which means it’s cool but not solving problems unless it is marketing and sales. I did not see any LegalTech tools! I guess lawyers were not the audience.

However, if you work in privacy, security, legal or governance at a large enterprise, then let me tell you the one thing that got my blood pressure spiking. AI agent creep.

Basically, AI agents are going to be so prolific that if you don’t get a grasp on them now, they will become your new “vendor management problem” and “shadow AI agent problem” combined.

It’s official. I give up. I might as well, because my voice is not being heard.

Which is ironic. For years, I have been warning about voice AI. It has a unique risk profile because a recording of your voice can send you to jail. It can also be used to fool people on the phone and lots and lots of malicious activities that cause real harm. I was particularly alarmed when voice was beginning to be used for verification and authentication when calling the bank.

Well, no one did anything about it and now it costs as little as a dollar, and a very short snippet of your voice (think that video you posted on social media), to clone your voice. I know my voice is out there. Eeek.

My advice: secret code words. Use them. Use them with everyone.

While in Jackson Hole visiting friends, their young daughters introduced me to the movie The Mitchells vs the Machines. I loved it, especially since it portrays a Silicon Valley tech “bro” who invents an AGI friend that he then offends, so the AI decides to eliminate all humans. It has a happy ending but it got me thinking about how this movie is perpetuating the fear and anxiety over AI. That, in turn, got me wondering, is it okay to create fear and anxiety about the dystopian possibility of AGI in a 5-year old?

Now, I don’t think either one of these two girls recognized the fear and anxiety the movie created because they both seemed pretty happy. And I suppose I could argue that many kids movies have a villain and a champion, so why is this different just because it uses AI as the villain vs a human? Which then gets me thinking that maybe this is Hollywood’s version of an AI governance program - using entertainment to teach kids that AI needs oversight.

Perhaps all AI governance training should require The Mitchells vs the Machines.