• by Steven Rosenbaum , Featured Contributor, 2 hours ago

I’ve been thinking a lot about robots lately, but not the ones that assemble cars or wander around warehouses. I’m thinking about the robots that call, email and text me every day, each one pretending there’s a person on the other end.

Every morning, I receive emails from various so-called people who claim to represent a family office or private equity firm. They are interested in investing in “my space,” whatever that might be. As evidence of their interest, they usually mention a company I founded many years ago or a place I worked many, many years ago.

Would I be available for a brief call to discuss an investment or possible buyout?

I get five or six of these emails a day. They follow a familiar sequence. The first one introduces the opportunity. A few days later, another arrives asking whether I saw the original email. Then comes the slightly more impatient follow-up. Each offers me an easy way to stop the messages: Just reply “not interested.”

Of course, replying would confirm that the email address works and that a human being is reading the messages. This is not exactly a new trick. What has changed is how relentless and inexpensive it has become.

The same thing is happening on the phone.

Once or twice a day, someone calls asking for Dr. Rosenbaum, the owner of a spine clinic, usually somewhere in Florida. Apparently, somebody bought a database of doctors or medical clinics and merged it with another list containing phone numbers and last names. Somewhere in that process, my number became attached to a medical practice.

When I explain that I am not a doctor, do not run a spine clinic and have no medical training, the caller often seems reluctant to believe me.

“Are you sure?”

Yes, I am reasonably certain that I am not a doctor.

Sometimes they ask whether I know another Dr. Rosenbaum, as if I might be able to provide a forwarding number. I don’t. But the calls have given me a glimpse of the staggering amount of automated junk that must pour into actual medical offices every day.

Then there are the political text messages.

“Hey Steve, it’s Chuck Schumer.”

Is it? I have no idea. It might be Chuck Schumer’s campaign. It might be a Democratic fundraising organization that has been authorized to use his name. It might be a completely unrelated political group. Or it might be a right-wing scammer impersonating Chuck Schumer to separate me from $25. “Chip In” they implore. Nope.

From the recipient’s perspective, these possibilities look almost identical.

That should trouble politicians. By adopting the language, frequency and emotional manipulation of scammers, legitimate campaigns have made themselves indistinguishable from fraud. Every message arrives with the same manufactured urgency. Democracy is at stake. A deadline is approaching. My donation will be matched by some improbable percentage, but only if I act in the next 17 minutes.

The campaigns may be real, but the communication feels fake.

All these messages share two things: bad data and almost nonexistent distribution costs. You might expect smarter technology to reduce the number of calls intended for a nonexistent Dr. Rosenbaum or emails offering to acquire a company I left years ago. Instead, the errors multiply because there is no meaningful penalty for making them.

AI will make this worse.

The emails are already becoming more persuasive. They address me by name and mention an article I wrote. Soon, they will include something about my neighborhood, my favorite deli, my dog or a family member. Each message will weave a few real facts into a fictional relationship.

The technology does not need to convince everyone. It only needs to make enough people pause before pressing delete.

That pause is the real product. Each message takes only a second or two of my attention, but the robots are indefatigable. They do not get tired, discouraged or embarrassed. They do not care that I ignored the first six messages. Sending the seventh costs essentially nothing.

The result is not merely annoyance. These robots are training us to distrust nearly every piece of inbound communication. A genuine investor, a real doctor’s office or an actual political campaign now has to overcome the suspicion created by millions of automated impostors.

Years ago, my friend Esther Dyson, author and technology thought leader, suggested putting a small cost on sending email. As I remember it, the amount was something like a penny. The point was not to make ordinary communication expensive. A penny would mean almost nothing to someone sending an email to a friend or colleague. But the economics change when a robot sends millions of messages.

Twenty unwanted emails would cost 20 cents. Two hundred would cost two dollars. A million would cost $10,000.

Suddenly, accuracy would matter. Senders might even have an incentive to determine whether I own a spine clinic before calling me about its billing system.

The details would be complicated. Who collects the money? Which messages are charged? How do we protect nonprofits, small publishers and people who legitimately communicate with large audiences?

But the underlying principle is right. If machines can communicate with unlimited numbers of people at virtually no cost, they will consume an unlimited amount of human attention. Blocking individual senders will not solve that. Spam filters will always be chasing the newest version. AI may help identify bad messages, but it will also help produce better ones.

The robots are not going away. So perhaps it is time to make them pay admission.

A penny may be almost nothing to a person. To a bad robot, it could be the beginning of accountability.