How to talk to shareholders about AI (how not to)

If a robot kicks in your front door, how will it come?

Just a few months after ChatGPT launched with the promise of making malicious writing reasonably accessible, many companies are beginning to come to terms with their obvious nemesis.

Although there are some twists

— Resistance to the learning models that underpin technology
— sheer bad input/output

The capital market does not waste time tearing clothes. According to the FT, last week:

Stocks in the education sector plummeted [on May 2nd] After earnings warnings from edtech firm Chegg, investors are betting on the potential of artificial intelligence to upend business models.

California-based online study guide provider Chegg acknowledged on Monday night that “significant student interest” in AI chatbot ChatGPT is beginning to hurt sales.

Director Dan Rosenweig said:

So this is not something that falls from the sky. This is just recognizing that there has been a technological change, and we need to adjust our company for it, be proactive about it, and adjust our cost structure accordingly. And we are doing all of that now.

Chegg’s stock price fell by about half by the end of the day. The predicament has been made even more grim by the fact that the company has been deploying his AI so fast (which Maine’s FT colleague Richard Waters has detailed here).

Other companies are now scrambling to avoid a similar scenario. Yesterday morning, educational group Pearson held a conference call with analysts and the media to discuss AI in particular (with its stock closing at about 1.75%), and RELX, which runs several information businesses, this afternoon. Hold a conference call.

ChatGPT’s success has sparked a lot of whining and “actually we did it first” quips from big tech companies, but Robin here about a study that quantified the hype impact of generative AI. I have written. Teacher, you are quite thick.

But tech giants aren’t a big deal at this point. It’s in the (relatively) welterweight category where AI can prove truly disruptive, and cautionary tales like Chegg’s suggest that smaller players are more proactive about the risks of AI. May encourage communication.

Citi’s European Media and Internet team explored this communication challenge as part of an effort by analysts to plan for stable AI risk for underlying stocks. Those people write:

Our conclusion is that, while generative AI promises to impact a significant proportion of the companies we cover, it likely poses as many opportunities as it does risks. is. We specifically focus on companies with domain-specific/unique skills/datasets and relationships with institutional customers, taking a contrarian stance on AI risk/reward.

When using the 4-input system:

Primary Impact: Short-term direct impact on AI.
Secondary impacts: direct long-term impacts of AI.
Cost impact: How will AI affect your company’s costs?
End-industry risk: Individually evaluate long-term impacts, including exposure to other industries that may be disruptive.

Here’s what they came up with:

SAES City:

So, this is somewhat arbitrary, but in short, we choose an AI risk score of 1.0 as what investors should consider.

Aggregated and graphed, it looks like this:

Expanding the analysis by company, Chegg reigns supreme as a very bad looking company with a score of 2.5.

Clearly for Chegg and those who hold shares in it, this doesn’t help at all as the price has already plummeted. But are there lessons to be learned regarding risk communication? what.

City:

Clearly “low impact” companies will likely need to do relatively little to address broader market-level concerns about potential disruption from generative AI-based technologies.

But there is some confusion for companies that may be more affected. Simply put, it is: Do people talk about it or not?

Analysts are blaming references to key AI-related buzzwords since November 1 and stock performance since March 14 in an attempt to make the connection.

From this point of view, Chegg is a bit like the obnoxious guy in the zombie apocalypse movie who keeps asking how badly it takes to get infected.

Analyst wrote:

The true correlation (and causation) is likely to be between the actual “AI perceived risk” and stock performance, but as long as the AI says it is a signal of “risk perception” , which clearly also indicates some relationship. But perhaps surprisingly, talking about this a lot doesn’t seem to have helped Chegg much. That said, this may indicate that some degree of aggressiveness can help.

This seems to use a lot of words to say “we don’t know.” And that’s before we realize the timeframe weirdness. Sure, ChatGPT-4 was released in his mid-March, but the constant “AI is coming for you” debate has been going on since (at least) late last year.

No doubt the great minds of the financial industry will be hard at work developing ways to stop their customers’ checks. Here’s what ChatGPT-4 recommends while you wait.