TED: AI is dangerous, but not for the reasons you think | Sasha Luccioni

The video discusses the current impacts of AI on society, focusing on sustainability, copyright, and bias. It highlights the environmental costs of AI models, the unauthorized use of artists' work for AI training, and the biases present in AI systems. The speaker emphasizes the importance of addressing these issues now, rather than focusing solely on future existential risks. Tools like Code Carbon and the Stable Bias Explorer are introduced to help measure and mitigate AI's impacts. The video concludes with a call to action for more transparency and informed decision-making in AI development and deployment.

What is a major concern about AI's impact on society?

• AI can reduce global warming.

• AI can make people more intelligent.

• AI can end humanity.

• AI can improve global communication.

What is a significant environmental concern related to AI?

• AI models are made of biodegradable materials.

• AI models consume vast amounts of energy.

• AI models reduce carbon emissions.

• AI models are powered by solar energy.

What tool was created to measure AI's carbon emissions?

• Green AI Monitor

• Carbon Footprint Analyzer

• Code Carbon

• AI Carbon Tracker

What issue do artists face with AI models?

• AI models help artists gain recognition.

• AI models provide free art supplies.

• AI models use their work without consent.

• AI models create new art styles.

What tool helps artists check if their work was used in AI training?

• Art Tracker

• Have I Been Trained?

• Creative Commons Checker

• AI Art Monitor

What is a common bias found in AI facial recognition systems?

• They do not recognize any faces.

• They are equally accurate for all skin tones.

• They are more accurate for women of color.

• They are less accurate for women of color.

What is the Stable Bias Explorer used for?

• To explore AI's environmental impact.

• To track AI's technological advancements.

• To analyze AI's economic benefits.

• To explore bias in image generation models.

What is a potential danger of biased AI models in law enforcement?

• They can lead to wrongful imprisonment.

• They can improve crime-solving accuracy.

• They can reduce crime rates.

• They can enhance public safety.

What is the speaker's view on focusing on AI's future risks?

• It is irrelevant to AI's progress.

• It is the only way to prevent harm.

• It distracts from current impacts.

• It is essential for AI development.

What is the speaker's final message about AI's development?

• AI development should be halted.

• AI will inevitably harm society.

• We can collectively shape AI's direction.

• AI's future is predetermined.