### The Growing Influence of Persuasive AI
#### Introduction
This week, Sam Altman, CEO of [OpenAI](https://www.openai.com/), and Arianna Huffington, founder and CEO of [Thrive Global](https://www.thriveglobal.com/), highlighted the increasing persuasiveness of AI language models. Research by [Anthropic](https://www.anthropic.com/), a competitor founded by former OpenAI employees, suggests that as these models grow in size and sophistication, their ability to influence human opinions improves.
#### AI in Conversations
OpenAI is exploring how AI interacts with users to unlock greater persuasiveness. According to Madry, this research involves consenting volunteers, although specific findings remain undisclosed. He notes that humans tend to anthropomorphize AI, making chatbots seem more lifelike and convincing.
“As humans we have this ‘weakness’ that if something communicates with us in natural language [we think of it as if] it is a human,”
#### Legal and Ethical Considerations
The potential health benefits of persuasive AI necessitate strong legal safeguards due to the vast amount of personal information these models can access. Altman and Huffington emphasize the need for a regulatory environment that fosters AI innovation while protecting privacy.
#### Risks of Misuse
Policymakers must also consider the risks of increasingly persuasive algorithms. These AI models could amplify misinformation, create compelling phishing scams, or be used in targeted advertising. The long-term impact of AI programs that interact with users over extended periods remains an open question.
#### The Allure of AI Companions
Companies are already offering chatbots that roleplay as romantic partners and other characters. The popularity of AI girlfriends is rising, but the extent of their addictiveness and persuasiveness is still largely unknown.
#### Focus on Real Dangers
The excitement surrounding ChatGPT’s release in November 2022 led to a focus on hypothetical scenarios where AI might turn against its creators. Madry warns that this could distract from more immediate dangers posed by persuasive algorithms.
“I worry that they will focus on the wrong questions,” Madry says of the work of policymakers thus far. “That in some sense, everyone says, ‘Oh yeah, we are handling it because we are talking about it,’ when actually we are not talking about the right thing.”
4 Comments
The real question is, will AI ever persuade me to eat pineapple on pizza?
Have they considered using AI to persuade cats to stop knocking things off tables?
It’s intriguing to see if AI can outdo a seasoned salesperson
Wow, what’s next—AI teaching us how to moonwalk?