“It always seems impossible until it is done."
– NELSON MANDELA
2022 is full of uncertainty and change. Personally, I was once a PhD student, then graduated at May, and became an entrepreneur.
I’m now building technologies for social good. Information Tracer has been my major focus.
In this post, I reflect and share why I keep building Information Tracer, and what lessons I’ve learned along the way.
I first built Information Tracer out of necessity — in 2020 I was studying how COVID-19 fake news spread on multiple social media platforms – including Twitter, Facebook, Reddit, and YouTube. I needed a tool to search for a URL, collect posts, and then generate summary statistics. I couldn’t find an easy-to-use tool at the time, so I built one from scratch.
It is one thing to build a tool for myself, it is another game to build a tool for someone else. Since graduation, I’ve rebuilt Information Tracer (several times!) to make it more powerful and user-friendly to a wider range of researchers.
As an entrepreneur, nothing is more satisfying when people use what you build, and find it useful. This moment came to me one day in 2021.
The story started with an email from Christina Taft, a researcher who was investigating what seems to be an online bot and defamation campaign against Amber Heard. Amber Heard and Johnny Depp are two Hollywood celebrities. They have been accusing each other of defamation for many years.
Christina used Information Tracer to trace URLs from multiple platforms, and discovered that some URLs that are negative toward Amber Heard are shared by accounts in a coordinated manner. For example, anti-Amber posts are created in a short amount of time, and with similar texts.
I found the problem interesting, and helped Christina on several data collection tasks. Christina eventually created a Github repo to document evidence of bot campaigns against Amber Heard. You can find cross-platform analysis based on Information Tracer data here. For me, it is a conviction that Information Tracer is a project worth pursuing.
In late 2021, I began to think more seriously about building Information Tracer after graduation. Two things become clear: (1) I need to find more customers to improve my tool (2) I need to find resources (money!) to sustain myself and eventually built a team.
An opportunity came when I received an email about the 1st NYC Media Lab AI and Local News Challenge. The goal is to bridge technologists with journalists. For me it was very appealing – fact checkers and investigative journalists can use Information Tracer to understand the spread of spam, misinformation and other suspicious patterns.
I applied for the Challenge with my advisor Professor Joshua Tucker. Fortunately we were selected as one of five start-ups. During the spring of 2021, all five teams met every week, shared our progress, interviewed potential customers, talked to industry leaders, and demoed our systems.
I am grateful for NYC Media Labs and especially Matt, our program manager. I find it helpful to be part of a “structure“ – be it a program, challenge, school curriculum. A good structure brings accountability, which ensures progress.
The Challenge ended in May 2022. That same month I graduated from NYU Data Science PhD program. I created a company called Safe Link Network to continue building cyber-safety technologies including Information Tracer.
My favorite quote is from legendary runner Eliud Kipchoge: “Only the disciplined ones are free in life. If you are undisciplined, you are a slave to your moods, you are a slave to your passions. That’s a fact.”
After graduation and the end of the NYU Media Lab Challenge, I was suddenly free. No thesis submission deadline, no demo preparation deadline, no customers to interview. Every morning I had to ask myself: what should I do, what’s next? Emotionally I became agitated. I started to question my own passion.
To navigate myself out of this challenge, I need to become more disciplined. I need to set goals, set deadlines, and set a routine that allows me to make progress.
I was in Cupertino at the time, where my girlfriend was doing her summer internship. I’ve been running for many years. South Cal has expansive low-elevation mountains, ideal for trail running and meditation. It is inside those mountains that I started to redevelop a routine.
Every morning, I will wake up, change, and drive to a nearby trail. My favorites are Rancho San Antonio Preserve. I will run 5km-15km depending on the condition. During my run, I will make an itinerary of what I need to do that day – applying for more challenges, reaching out to potential customers on Linkedin, adding features to Information Tracer, reading a book chapter, etc,.Once I’m back home, I’m ready to start the day.
During the summer I wrote many emails and applications. Most went unanswered. Eventually, I heard one positive reply – I was invited to present my research at DEFCON Las Vegas. This rekindled my hope, and prompted me to prepare to demo for DEFCON.
DEFCON is one of the world’s largest hacking conferences. I gave a talk at Misinfo Village, where the focus is detecting online misinformation and abuse.
During DEFCON, I joined a 3-day “hacker running club”, and did a solo trail run at Red Rock Canyon National Conservation Area (pro tip: Turtlehead Peak is easier running up than down). Why change your good routine even at a conference in Las Vegas?
I came back to New York City in September. I followed up with people I met in DEFCON. We had great conversations, but eventually our discussions lost steam. People are busy with other tasks. Establishing a partnership with Information Tracer is not any company’s top priority.
I was back to where I stood right after graduation. Even though I spent a few more months developing Information Tracer, I could not find another customer. Reality told me to give up, and put the project on hold.
Once again, something unexpected happened. This time, it is an email from Twitter, with a title “Congratulations…”.
In August 2022, I found Twitter Chirp Developer Challenge, and submitted Information Tracer to contest under the Content Safety Tools category. In October, the Twitter Developer Platform Team informed me that I’m the winner. We signed a contract, in which Twitter agreed to launch a free promotion campaign for my technology. The team interviewed me and hired a professional designer to make edits. After the interview, I sent Twitter an extensive list of writing materials about Information Tracer, why I made it, what motivated me, etc,.
According to the plan, Twitter will hold a Developer Conference in San Francisco on November 16th. A director will announce winners at the conference. They will also launch the promotion campaign.
Then Elon Musk acquired Twitter. First we received an email that the Conference was simply canceled. Several days later, my contacts at Twitter were gone. Either they were fired, or they left themselves – I am not sure. Emails were bounced back. This is sad because the Twitter Developer Platform Team has been instrumental for researchers who analyze social networks for public good.
I managed to email one Twitter employee, who told me they are “operating with a very lean team. So while I do not have enough context, I will try and get information for you on this campaign.” No update so far…
The “old Twitter” cared about conversation safety, and wanted to help promote tools that track online misinformation. The “new Twitter” has deprioritized conversation safety.
After the Twitter saga, I feel I lost my sight again.
This time however, I decided to be more proactive. If I can’t find an existing opportunity, I will create one.
Currently, I’m working with my friends to jointly apply for a CDC grant to develop a system to predict vaccine misinformation. If accepted, Information Tracer can help health agencies to promote true information, and ultimately drive up the vaccination rate.
I’m also discussing with a few folks on how to use Information Tracer to help humanitarian workers. Many frontlines need a robust system for information awareness. Disinformation is rampant during humanitarian crises.
In particular, frontlines need to understand (1) what is being discussed about the crisis (2) if social media platforms are being weaponized (3) who are spreading spam, fake news or disinformation. Information Tracer can be a powerful tool in the arsenal of humanitarian technologies.
Many discussions are in early stages. I believe if we have a right mission, and we do the right thing, we will have the right outcome.
Zhouhan,
New York City,
December 28, 2022
p.s.: If you have any thoughts or ideas, whether related to Information Tracer or not, let me know. Always happy to chat. (email zhouhan[dot]chen@nyu[dot]edu)