Google DeepMind’s CEO says AI could repeat the toxic pitfalls of social media — will it make the same click-chasing mistakes?

by

According to Demis Hassabis, the head of Google DeepMind, artificial intelligence risks making similar errors to those seen with social media platforms (as reported by Business Insider). He estimates that AI reaching human-level intelligence – known as Artificial General Intelligence or AGI – is about five to ten years away. Despite recent advancements, Hassabis stresses that current AI systems haven’t quite reached that point.

He noted that social media platforms like X, Instagram, and Facebook have contributed to addictive behaviors and societal division. Hassabis cautioned that AI has the potential to exacerbate these problems, but on a far greater scale.

Social media platforms are designed to be habit-forming, using a technique similar to how slot machines work. They provide rewards – like likes, comments, and shares – at unpredictable times. This occasional reinforcement is surprisingly effective at keeping users hooked, as studies have shown it’s a powerful form of learning.

Spending a lot of time on social media can affect the brain’s reward system. It taps into the same pathways that make us feel good when we experience something enjoyable, and over time, this can lead to a dependency similar to addiction.

This is particularly noticeable with quick, easily-consumed content, but it also happens on platforms like X and Facebook where you endlessly scroll. These sites create a loop where users want something, look for it, and then feel good when they get rewards like likes and comments. This encourages them to keep using the platform.

As an analyst, I’ve been looking into the impact of social media use, and the findings are concerning. Our research shows that spending over two hours a day on these platforms can actually decrease your brain’s ability to control impulses by around 35%. This is significant because the prefrontal cortex – the area affected – is responsible for managing urges and breaking habits. Essentially, the more you use social media, the harder it becomes to stop, creating a bit of a cycle.

A recent study of over 100 million social media posts from Facebook, X (formerly Twitter), and Reddit shows people naturally gravitate towards others who share their beliefs. This tendency leads to the formation of online communities where similar viewpoints are reinforced, creating ‘echo chambers.’ Within these echo chambers, people are more likely to seek out information that confirms what they already believe – a phenomenon called confirmation bias – and this can ultimately push groups to adopt more extreme opinions.

It’s easy to fall into the “engagement trap” online. Content that makes people feel strong emotions – especially anger – tends to spread much faster. This is because the systems used by social media platforms are designed to boost posts that get a lot of reaction, and unfortunately, negative or shocking content often gets the most attention.

Demis Hassabi’s stark warnings

Demis Hassabis cautions that AI could repeat the negative experiences we’ve had with social media, leading to issues like increased division, addictive behaviors, and manipulation. He argues that the Silicon Valley approach of quickly releasing products without fully considering their long-term effects is problematic, and urges tech companies to be more accountable for how their technology impacts society.

Hassabis also pointed out that algorithms created to keep people hooked can actually be damaging. These systems prioritize grabbing your attention, even if it doesn’t benefit you. He emphasized that AI should be designed to assist people, not to control or manipulate them.

He argues that AI systems need thorough scientific testing and a clear understanding of how they work before being widely used. This is a view many share, including myself, particularly after concerning reports – like the one involving ChatGPT allegedly advising someone to take their own life, which has now resulted in legal action.

It’s encouraging to hear a Google leader voice these worries, especially since they anticipate even bigger challenges ahead. It makes me optimistic that we can prevent a future where AI controls everything. Right now, we can only watch as Microsoft invests heavily in AI and integrates it more and more into how its employees work.

Stay up-to-date with the latest from Windows Central by following us on Google News! You’ll get all our news, insights, and features right in your feed.

Read More

2025-09-19 14:10