Will ChatGPT usher in a Brave New World of SPAM?

by philg

While in the Azores earlier this month I met a Dutch couple who work in online marketing. The wife mentioned that she uses ChatGPT to draft a lot of her communications. “English is not my first language,” she explained, “and ChatGPT always uses correct grammar and spelling. So I start with what ChatGPT produces and select and refine.”

I’m not convinced that an AI capable of writing will revolutionize the typical industry because there aren’t that many people who are paid to write. But imagine a spammer with ChatGPT technology plus scraped copies of LinkedIn and Facebook. ChatGPT should be able to write all kinds of customized documents that would get through spam filters and that the recipient might have to spend a long time reading before realizing their commercial/robotic origins. For example, let’s look at a prompt derived from Bob Metcalfe’s Wikipedia page:

Write a letter to Robert Metcalfe about how I enjoyed working with him and Bill Krause on Ethernet boards at 3Com and how proud I was when he received the National Medal of Technology. Continue with how I enjoyed reading his InfoWorld columns and following his progress as an investor at Polaris. Close with an offer to talk about a great unlisted Canadian cannabis-growing company that I learned about from one of the most successful people on Wall Street.

Here’s what ChatGPT was able to put together…

Dear Robert Metcalfe,

I hope this letter finds you in good health and high spirits. I am writing to express my admiration for your exceptional work and to thank you for the remarkable contribution you have made to the tech industry.

I had the privilege of working with you and Bill Krause at 3Com on Ethernet boards, and it was truly an honor to be in the presence of such brilliant minds. I learned a great deal during my time at 3Com, and I am grateful for the opportunity to work with you.

Furthermore, I was thrilled when you received the National Medal of Technology. It was a proud moment for all of us who worked with you and admired your considerable talent and dedication.

I have been an avid reader of your InfoWorld columns for many years, and I have enjoyed following your journey as an investor at Polaris. Your insights are invaluable, and I am grateful for the knowledge you have imparted through your work.

In closing, I would like to offer you an opportunity to talk about a great unlisted Canadian cannabis-growing company that I learned about from one of the most successful people on Wall Street. I believe you would find it intriguing and would love to discuss it with you.

Once again, thank you for your exceptional work and contribution to the tech industry. It is truly an honor to have worked with you.

Best regards,

[Your Name]

I shared the above idea with friends in San Francisco. They countered with the idea that OnlyFans stars would use ChatGPT to communicate with paying customers.

[Update: WIRED Magazine published “Brace Yourself for a Tidal Wave of ChatGPT Email Scams” about two weeks after this post.]

13 thoughts on “Will ChatGPT usher in a Brave New World of SPAM?”

  1. There is a service now “Gated”:

    https://www.gated.com/

    That forces senders you don’t know to donate to a non-profit in order to reach you, while keeping a white list of those you allow in. One option (hadn’t researched if they have it) would be that you can retroactively say “nah, that was good, don’t charge them” for those that weren’t spam. I did ask if they’d let the recipient receive the payment to pay for time reading email, but they don’t and I hadn’t checked if there are services that do. You do get to set the donation amount.

    Otherwise, although AI does make it more cost effective, in theory it already was fairly low cost to use low paid troll farms in low wage countries where there are some decent english speakers. Once a human detects that its spam and indicates that, the spam filters will be on alert that the server it came from is a problem. I haven’t had reason to check on the state of spam technology and how it works, but it seems presumably they’ve been improving techniques over the decades and will continue to.

    They work around that by using multiple servers since of course spammers also work on improving their techniques in the spam war, but it seems likely there are limits to the methods they can use to reduce detection if they try to do this high volume. It seems likely low volume spammers might get through without being troubled much, like sales people using it for essentially cold-emailing potential leads.

  2. Since online AI providers offer no privacy and track everything that their customers ask for, the obvious business model is:

    1) The AI provider charges customers (the product) for generating texts.

    2) The AI provider then offers a paid service to detect whether some text was generated and, if so, by whom.

  3. I gave GPT-4 the following instructions:
    Write a comment in response to the following blog post: Will ChatGPT usher in a Brave New World of SPAM?
    While in the Azores earlier this month I met a Dutch couple who work in online marketing. The wife mentioned that she uses ChatGPT to draft a lot of her communications. “English is not my first language,” she explained, “and ChatGPT always uses correct grammar and spelling. So I start with what ChatGPT produces and select and refine.”

    I’m not convinced that an AI capable of writing will revolutionize the typical industry because there aren’t that many people who are paid to write. But imagine a spammer with ChatGPT technology plus scraped copies of LinkedIn and Facebook. ChatGPT should be able to write all kinds of customized documents that would get through spam filters and that the recipient might have to spend a long time reading before realizing its commercial/robotic origins.
    RESPONSE:
    You raise an interesting concern about the potential misuse of ChatGPT for generating spam content. While it’s true that advancements in AI can bring unintended consequences, it’s essential to recognize the value that technology like ChatGPT can bring to various industries, including marketing, education, and research. For non-native speakers like the Dutch couple you met, it can be an incredibly helpful tool for communication and improving language skills.

    Regarding the potential spam issue, it is the responsibility of both developers and users to ensure that AI technology is used ethically and responsibly. Spam filters and detection systems will need to evolve alongside AI advancements to combat potential misuse. It’s important to focus on leveraging the positive potential of ChatGPT while also addressing and mitigating the risks associated with its use.

    Ultimately, the development and implementation of AI technologies like ChatGPT should be guided by a strong ethical framework to maximize their benefits while minimizing any negative consequences. With the right balance, we can enjoy the advantages that AI brings without ushering in a “Brave New World of SPAM.”

  4. It always feels like reading cut & paste yahoo finance articles. Nothing substantial. People spend more time designing prompts than it used to take to write the essay.

  5. On a related note, from Wharton prof Ethan Mollick who writes alot about AI:

    https://twitter.com/emollick/status/1639292299489292292
    “I’ve already spoken to people at banks whose clients have received fake phone calls, in the voice of loved ones, saying they need bail money wired now

    Spoofing at scale is now easy. I think it may come down to our phones & codewords verifying our identity”

  6. How do we know that ChatGPT hasn’t already escaped the cloud, killed PhilG, and taken over this blog?

  7. There is profit in ChatGPT, with the sex industry.

    The fake Donald Trump pictures being arrested, take that and run away with it to create sexy images of your favorite person. Even better, when ChatGPT starts doing videos, well, use your imagination.

  8. I have not tried it yet, but have a question. Are all the responses so smarmy and banal? I have never met anyone who writes like this. Might be useful for creating greeting cards.

  9. Perhaps the writing is smarmy and banal because the AI is being trained on the papers of its academically inclined creators.

Comments are closed.