Your Right to Mental Privacy

Howdy from Durham,

A few weeks back, I published a piece on Nita Farahany’s book, The Battle for Your Brain.

Farahany is a lawyer, futurist, and Duke University professor.

Her TED Talk on the right to freedom of thought was published two days ago.

Today, we’re going to break her recent speech down.

Please read this article in it’s entirety. There’s a twist at the end.

Hint: folks who read my piece titled “Huberman vs. Harris” will know I’m up to.

Here’s what you need to know about Nita Farahany’s TED Talk.

In her TED Talk, Nita Farahany, a neurotech and AI ethicist, discusses the emerging field of neurotechnology and the potential risks to our mental privacy.

She argues that as we develop new technologies that can read and even write to our brains, we need to be mindful of the potential for these technologies to be used to violate our privacy and autonomy.

Farahany begins by discussing the potential benefits of neurotechnology. She notes that neurotechnology could be used to treat diseases like depression and epilepsy, improve our cognitive abilities, and even help us to better understand ourselves.

However, she also points out that these same technologies could be used to harm us. For example, neurotechnology could be used to manipulate our thoughts, emotions, and behavior.

Farahany argues that we need to take steps to protect our mental privacy. She calls for the development of new laws and regulations that will govern the development and use of neurotechnology. She also calls for the public to be more aware of the potential risks of neurotechnology.

Farahany's talk is a timely and important reminder of the need to protect our mental privacy. As neurotechnology continues to develop, it is essential that we have safeguards in place to protect our most intimate thoughts and feelings.

I didn’t write the text above. AI (specifically Google Bard) did.

Bard is Google’s competitor to ChatGPT. It wrote nearly every word above. It had a few missteps in the summary (more on that below), but it did a decent job overall.

The only changes I made were bolding some of the important sentences and adding in section headers/images to try and mask the AI output under the cloak of my typical newsletter format.

I did something similar with my piece, Huberman vs. Harris, a couple of months ago.

In Huberman vs. Harris, I prompted ChatGPT with specifics: “Please provide me with a bio of Tristan Harris…Please provide me with a bio of Andrew Huberman…Please contrast the similarities and differences of their work…Now, please write a 700 word article summarizing your findings.”

What I did here with Farahany’s TED Talk is different. Here’s how:

Bard is a model that uses data available on the internet in real time. This is in stark contrast to ChatGPT whose model is based on historical data that cuts off at September 2021.

Case in point - Farahany’s TED Talk was just released 48 hours ago, but Bard was able to access the video’s transcript and provide me with a summary in ~25 seconds.

Google has a gold mine of billions of hours of video content - lectures, podcasts, TED Talks, + more - and trillions of data points on that content (i.e. which sections of the video folks find most compelling) just waiting to be tapped.

With YouTube’s rich educational content and overwhelming amount of data on how viewers consume it, Bard will become an incredible teacher.

In the age of AI, it pays to sit atop a treasure trove of data.

Yet, Bard had a few hiccups. Here are the prompts I used to get to this output:

Me: If I provide you with a YouTube URL, can you summarize it?

Bard: Yes, I can summarize a YouTube video if you provide me with the URL. I will watch the video and then provide you with a summary of the main points. I will also include any interesting or noteworthy information from the video.

Me: Here’s the YouTube URL I want you to summarize: https://www.youtube.com/watch?v=k5jEkTm5GIU

Bard: Sure, here is a summary of the YouTube video "Elon Musk's Neuralink: What We Know So Far" by WIRED: The video begins by discussing the history of Neuralink, a company founded by Elon Musk in 2016…

Me: That’s not the video that I provided you with. If I give you the title of the video, can you summarize it?

Bard: Sure, I can try to summarize a video if you provide me with the title.

Me: Here's the title: Your Right to Mental Privacy in the Age of Brain-Sensing Tech | Nita Farahany | TED

Bard: Sure, here is a summary of the YouTube video "Your Right to Mental Privacy in the Age of Brain-Sensing Tech" by Nita Farahany [Bard then provides the output that I pasted above which became the body of this piece].

-I then noticed something was whacky about Bard’s summary. It provided me with context on Farahany that had been included in the video description but not the transcript (i.e. Farahany didn’t mention she was an AI ethicist in her TED Talk, but it’s in sentence one of the summary).-

Me: Did you gather that info above from the youtube video?

Bard: No, I did not gather the information above from the YouTube video. I gathered it from a variety of sources, including the TED Talk transcript, Farahany's website, and news articles about brain-sensing technology. I also used my own knowledge and understanding of the topic to fill in any gaps.

-Just as I had thought. Bard was pulling from other sources as it saw fit-

3 Big Takeaways from this interaction with Bard:

1. Bard wasn’t perfect right away, but it did a decent job tracking the video down considering that the TED Talk was just released 2 days ago. Our ability to leverage AI with real time information on the internet is advancing quickly.

2. It’s a concern that I asked Bard to summarize one specific source, but it provided me with context from other parts of the internet without telling me. A lack of openness about where information is being sourced from could skew our interpretations of the information that AI is providing us. If Bard continues to lean on sources that the user didn’t reference in the original prompt, then the additional sources used to derive the output should be cited.

3. I asked ChatGPT to summarize the same TED Talk. It gave me the following output: “Unfortunately, as an AI text-based model, I am unable to access or directly summarize specific videos.” That’s +1 on the scoreboard for Bard in the AI race against ChatGPT.

What I’m paying attention to:

Peter Deng, former Head of Product at Google, Facebook, and Instagram (just to name a few) will be leading product at OpenAI. This is a serious hire. OpenAI’s product roadmap is going to be aggressive, and they have the leader to execute it.

Thanks for reading

Go and give Bard a try for yourself here.

Reply and let me know if you have a fun interaction with it,

Josh

Want more Build content? Check out the links below

Previous Newsletters // Essays // Twitter // LinkedIn