Navigating Permission in the Age of AI
The Consent Conundrum

Imagine coming home to find everything you love—ice cream, book recommendations, even the perfect birthday gift—waiting for you. This isn’t the end of a romcom, but your smartphone’s AI assistant. Welcome to the wild world of AI consent, where checking 'I Agree' isn't just about lending your email address anymore—it's about handing over the keys to your digital life.

By Nidhi Singh

Remember when consent was as simple as checking a box that said "I agree" (which, let's be honest, none of us ever read)? Well, buckle up, because artificial intelligence is about to make that little checkbox look as outdated as a feature phone at a tech convention.

The Changing Model of Consent

Here's a sobering thought: Nobody asks bystanders to sign a consent form before they get hit by a self-driving car. While the driver signs plenty of liability waivers when purchasing their autonomous vehicle, the pedestrians who might be affected never get a say. This highlights a fundamental problem with AI consent - it's not just about individual choice anymore; it's about societal impact.

Traditionally, digital consent has been pretty straightforward: you agree to share specific pieces of information, like your email address or location, for specific purposes. It's like letting someone borrow your car - you hand over the keys, set some ground rules ("Don't mess with my Spotify settings!"), and that's that.
But here's where things get wild: Imagine if instead of borrowing your car, someone borrows your entire lifestyle—likes, dislikes, and everything in between. That's essentially what happens when you interact with AI systems today.

‘Consenting’ to the use of AI is far more complicated and nuanced than ever before and none of our old consent mechanisms fit the bill anymore. Experts are now breaking down AI consent into the “Three C's":

1. Context: Understanding when and where your data can be used, and by whom
2. Consent: Being properly informed about how that data will be used
3. Control: Having the ability to refuse certain uses of your data

The Butler Did It: Manipulations of the AI Chatbot

Let's meet Simran (completely made up, but probably just like someone you know). With offers from the annual Diwali sale lighting up every screen in her life, she finally decided to purchase a speaker with a voice assistant this year. Now she can groove to Bollywood music in her kitchen, call her parents, order online, set reminders and send text messages with her voice alone. However, at the same time, she has also invited an AI system into her home which:

• learns her daily routine,
• understands her music taste, shopping habits, and even her mood patterns,
• can interact with her shopping account, digital wallets, read her texts and make phone calls,
• knows when she’s been bad or good.

But here’s the crux of the arguments: Simran consented to all of this—The one tick in the checkbox and a quick "Yes" to the setup guide. But how can anyone truly consent to something they don't fully understand?

Here’s where things take a fascinating turn. Modern AI isn't just collecting data—it's learning from it and adapting its behaviour to influence yours. Consider these totally-possible-but-slightly-scary scenarios:

1. Your AI assistant notices you tend to impulse shop when you're stressed. Suddenly, those targeted ads start appearing right after your weekly team meetings.
2. Your smart home system learns that you're prone to impulse purchases in the morning after coffee. Now, your shopping app doubles up on ads early in the morning.
3. Your AI chatbot observes a monthly recurring lunch with your extended family on the last Sunday. It now shows you constant ads of dating apps and matrimonial websites as you’re more susceptible to signing up for such services after these gatherings.

Congratulations! Your personal AI chat companion has learned exactly which emotional buttons to push to keep you engaged in conversation.

An Impossible Permission Slip

Here's the real mind-bender: You might be reading this article right now because an AI algorithm determined you'd be interested in it based on your browsing history. And it was probably right, wasn't it? Welcome to the future - where consent is complicated, privacy is a puzzle, and we're all trying to figure out how to navigate the brave new world without accidentally giving our AI assistants too much power over our lives.

Sometimes, the developers of AI systems cannot fully explain how their AI reaches specific decisions. This "black box" nature of AI systems—where developers and users are unable to understand the actual decision making process of an AI system—makes informed consent particularly challenging and raises a crucial question: How can users meaningfully consent to something that even its creators don't fully understand?

So, here's the million-dollar question: How can you meaningfully consent to something that:

• evolves and learns over time,
• can predict (and potentially influence) your behaviour,
• understands you better than you might understand yourself,
• has implications that even its developers might not fully comprehend,
• affects not just you, but society as a whole.

A simple answer is—you can’t, at least not in any meaningful way. Our traditional notion of consent needs an upgrade. We're trying to use an idea which was fundamentally for simple yes/no decisions in a world of complex, evolving AI relationships. It's like trying to perform brain surgery with safety scissors - the tool just isn't up to the task.

The future of consent might involve:

• dynamic permissions that evolve with the AI's capabilities,
• regular "consent check-ins" to help you understand how the AI is using your data,
• clearer boundaries for acceptable behavioural influence,
• the right to "pull the plug" and truly disconnect (if that's even possible anymore),
• societal-level consent mechanisms for technologies that affect everyone,
• preference signals for creators to control how their data is used in AI training.

The Societal Conversation We Need to Have

As one expert noted, "Society needs to have a conversation about technology." Though these conversations happen in fragmented forms, we need a society-wide dialogue because everyone is impacted by these new technologies.

The days of "it's better to ask forgiveness than permission" in tech development must end. As AI becomes more powerful and is applied to more fields, it’s essential to build frameworks for both individual and societal consent. We need to shape a future that leverages tech for societal progress, not just consumer impulses—like manipulating us to spend more money on a new magic bullet blender (that you don’t need) to make healthy smoothies (which you won’t drink).

Don't tell your smart home assistant I said that. It might get ideas!