AI agents will outperform salespeople by improving persuasion
AI agents will outperform salespeople by improving persuasion
By Louis Rosenberg | Published: 2025-02-17 17:31:00 | Source: The Future – Big Think
Sign up for Big Think on Substack
The most surprising and impactful new stories delivered to your inbox every week for free.
When I was a kid, there were four AI agents in my life. Their names were Inky, Blinky, Pinky and Clyde and they did their best to chase me. It was 1980, and the agents were the four colorful ghosts in the popular arcade game Pac-Man.
By today’s standards, they were not particularly intelligent, yet they seemed to be pursuing me with cunning and intent. This was decades before neural networks were used in video games, so its behaviors were controlled by simple algorithms called heuristics that dictated how I was chased through the maze.
Most people don’t realize it, but each of the four ghosts is programmed with a different “personality.” Good players can monitor their actions and learn to predict their behaviors. For example, the red ghost (Blinky) is designed with a “stalker” character heading straight towards you. The Pink Ghost (Pinky), on the other hand, is given an “ambush” personality that predicts where she’s going and tries to get there first. As a result, if you lunge directly at Pinkie, you can use her Personality against her, causing her to move away from you.
I remember that in 1980, a skilled human could observe these AI agents, decipher their unique personalities, and use those insights to defeat them. Now, 45 years later, the tides are about to turn. Like it or not, AI agents tasked with decryption will soon be deployed Your personality So they can use those ideas to Optimum effect You.
In other words, we are all about to become unwitting players in the “human game” and the AI agents will be the ones trying to get the highest score. I mean this literally: most AI systems are designed to maximize a “reward function” that earns points for achieving goals. This allows AI systems to quickly find optimal solutions. Unfortunately, without regulatory protection, we as humans are likely to become the goal That AI agents are tasked with improving.
“The problem of artificial intelligence manipulation”
I’m very worried about Conversational agents It would engage us in friendly dialogue throughout our daily lives. They will speak to us through real-life avatars on our computers and phones, and soon, through AI-powered glasses that guide us through our days. Unless there are clear limitations, these agents will be designed to probe us via conversation for information so that they can describe our moods, inclinations, personalities, and desires, and use those traits to Maximize their persuasive effect When selling products to us, providing services to us, or convincing us to believe false information.
This is called “The problem of artificial intelligence manipulation“I have been warning about the risks for years. So far, policymakers have not taken decisive action, viewing the threat as too far in the future. But now, with the release of Deepseek-R1, the final hurdle to the widespread deployment of AI agents, Real-time processing costhas been significantly reduced. Before this year is out, they will become AI agents A new form of targeted media This is the case Interactive and adaptiveIt will improve its ability to influence our thoughts, direct our feelings, and direct our behaviors.
Of course, human salespeople are reactive and adaptive as well. They engage us in friendly dialogue to size us up, and they quickly find what buttons they can push to influence us. AI agents will make them look like amateurs, capable of subtly extracting information from us, which might intimidate an experienced therapist. They will use these insights to adjust their conversational styles in real time, working to persuade us more effectively than any used car salesman.
It will be these Asymmetric encounters Where the artificial factor has the upper hand. After all, when you’re dealing with someone who’s trying to influence you, you can usually sense their motives and sincerity. It won’t be a fair fight with AI agents. They will be able to inflate your size with superhuman skill, but you will not be able to inflate them at all. That’s because they will look, sound, and act so human that we will Trust them unconsciously When they smile with sympathy and understanding, they forget that their influence is just an illusion.
Additionally, their voice, vocabulary, speaking style, age, gender, ethnicity, and facial features are likely to be tailored to each of us personally. Maximize our receptivity. Unlike human salespeople who need to size up each customer from scratch, these virtual entities have access to stored data about our backgrounds and interests. They can then use this personal data to quickly gain your trust, asking you about your kids, your job, or perhaps your favorite Yankees, making it easier for you to unconsciously let your guard down.
To educate policymakers about the dangers of manipulation using AI, I helped produce an award-winning short film called Privacy lost It is produced by Responsible Metaverse Alliance, Minderoo Pictures, and XR Guild. the 3-minute narration It depicts a young family eating in a restaurant while wearing augmented reality glasses. Instead of human servers, avatars take each diner’s orders, using the power of artificial intelligence to sell them in personalized ways. The film was considered science fiction when it was released in 2023, but just two years later, major tech companies are getting involved in a full-scale operation. The arms race to make artificial intelligence-powered glasses Which can be misused in these ways.
I’m not a fan of overly aggressive regulation, but we need smart, tight restrictions on AI to avoid that.Superhuman manipulation“By conversational agents. Without protection, these agents will convince us to buy things we don’t need, believe things that aren’t true, and accept things that are not in our best interests. It’s easy to tell yourself that you won’t be vulnerable, but with AI improving every word they say, we’ll likely all be outdone.”
One solution is to prevent AI agents from creating “Feedback loops“They improve their ability to persuade by analyzing our reactions and repeatedly adjusting their tactics. Additionally, AI agents should be asked to inform you of their goals. If their goal is to persuade you to buy a car, vote for a politician, or pressure your family doctor for a new drug, these goals should be stated upfront. Finally, AI agents should not access personal data about your background, interests, or personality if that data could be used to influence you.”
In today’s world, target impact is a huge problem, and is often deployed as a gunshot fired in your general direction. Interactive AI agents will transform the target effect into… Thermal missiles That finds the best way to each of us. If we do not protect ourselves from this danger, I fear we will all lose the human game.
Sign up for Big Think on Substack
The most surprising and impactful new stories delivered to your inbox every week for free.
ــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــــ
