Alecto AI

Digital abuse such as revenge porn is extremely destructive for its victims, it can kill. Platforms rely on AI tools to fight digital abuse. However, AI industry has long been dominated by white males and is mostly institutionally focused instead of human-centered. On that basis, existing facial recognition tools have a lower accuracy rate when it comes to recognizing faces of women or people of color. 

Initiated by a revenge porn survivor, built by women of color, Alecto aims to end both image-based abuse and racist facial recognition with compassion and cutting-edge technology. We offer a biometrical protected, antiracist facial recognition tool that allows individuals to quickly search for infringing images of themselves online. 

Alecto is beyond a product, it’s a movement against digital abuse and algorithmic biases. We are striving towards a future where no one will ever have to suffer from algorithmic injustice or digital abuse. 

Alecto can solve 2 major issues altogether with one product: online image abuse (revenge porn, impersonation, cyberbullying), also biased facial recognition. Research shows that 1 in 12 people have been victims of revenge porn at some point during their lives, and 1/3 of them have considered suicide. 

For revenge porn victims, damage control is particularly painful. The infringing contents are hosted on various different platforms, it is difficult to search for these contents scattered all over the internet, they are also forced to relive the trauma over and over again while doing so. 

We need to be proactive when it comes to defending our privacy. But we can’t defend ourselves unless we have access to technology that can help us do so. 

The most efficient way to tackle online image abuse is through facial recognition. However, according to an MIT study, the data set used to train the facial recognition machine learning model is more than 77 percent male and more than 83 percent white. Examination of facial-analysis software shows an error rate of 0.8 percent for light-skinned men, but 34.7 percent for dark-skinned women. With 75% of revenge porn victims being female, it calls for an unbiased, human-centered solution. 

Alecto is an unbiased facial recognition tool accessible for individuals to quickly search for infringing images online. When it comes to privacy, people want to make sure the tools they use to defend themselves can’t be abused by any random third party. So we enabled biometric verification on Alecto, making sure that only you can access the sensitive facial information of your own. With the great gamification feature built into the app, we also allow you to constantly train the machine learning model to recognize your face better over time, minimizing the risks of algorithm bias based on race or gender. To add an extra layer of security, we do not save any data on our server, and we use E2EE to protect your data, everything will be automatically erased after each training session. 

To better deliver compassionate messages to all users and bring social values to the masses, we are aiming at creating a community through our app and multiple social media channels. In this caring online environment, Alecto will be able to provide useful legal resources, helps for mental health and spiritual healing. We also encourage individuals to explore their deepest thoughts and feelings by sharing their stories. 

Alecto will directly improve the lives of image abuse victims and people of color. To have a better understanding of what inspired the idea of Alecto, please read our founder’s compelling story as a revenge porn survivor: https://alectoai.medium.com/privacy-the-price-we-pay-for-technology-my-story-as-a-revenge-porn-survivor-cbb1a6fedb77 (This article is an important part of our application, please be sure to read it) 

The team lead of Alecto is a revenge porn survivor and a woman of color. Her personal experience helped her understand the pain points and needs of Alecto’s target population. 

Comparing with existing facial recognition companies such as Clearview AI, which has long been criticized for providing racially and sexually biased facial recognition to law enforcement, and resulted in multiple incidents of wrongful arrest of people of color; Alecto provides an alternative approach to address algorithm bias -- by putting the control back into people’s own hands. 

Alecto was designed to address both online image abuse issues and racial/sexually biased facial recognition. By offering a compassionate, unbiased facial recognition tool that helps online image abuse victims to defend their privacy, we are making the online environment safer and more friendly, not only for communities of color, but also for everyone. By allowing individuals to safely train the machine learning model without jeopardizing their privacy, we are also making the facial recognition algorithm more human-centered, therefore minimizing the risks of racial or gender bias.  

Our prototype is currently being tested within a small group of people on the algorithm's stability and error rate. Additionally, we are making improvements on user experience during this stage. Our beta test is scheduled to happen in August, which will allow more people to test out the product before the official launch. 

Alecto resolves 3 major issues with 1 solution, as shown above. 

Biased facial recognition is the reflection of an institutional-focused AI industry. The cause of algorithm bias is due to the lack of enough training data from people of color and women, the solution to this issue is simple: acquire more training data from these marginalized groups. Almost all existing facial recognition solutions scrape the public database (without asking for permission) for training data. Our approach empowers individuals to control their own privacy, we use the data provided by individuals (willingly) to train the ML model for best results. Alecto is compassionate and human-centered, it’s beyond a product, it is the guardian angel of individuals’ privacy. 

As we scale to the global level, Alecto will not only end revenge porn or other forms of image-based abuse, we can also prevent biased facial recognition abuse. Clearview AI’s algorithm resulted in multiple incidents of wrongful arrest and imprisonment of people of color, such abuse could be avoided with Alecto’s approach. 

We are currently at the prototype stage and are still testing the product for its stability and accuracy rate within a small group of people, so we are not serving anyone publicly due to privacy and safety concerns. 

In one year, we plan to reach 500k registered users and 50k weekly active users. 

In 5 years, we plan to reach 20 million registered users and 2 million daily active users. 

One of the most important impact goals is to help more people with Alecto, and the metric to measure that is user growth and user experience. Some of the quantitative metrics include user growth rate, registered users, DAU, WAU, MAU, etc.

For qualitative metrics such as user experience, we will have a professional UX research team to interview focused groups in order to listen to user feedback and make improvements 

For technology impact, we will compare our algorithm error rate based on race and gender with other industry solutions and see how much we could outperform them. 

For social impact, we will look at how many states/countries have criminalized revenge porn. 

3 full time, 2 part time 

Our CEO is a revenge porn survivor, Due to privacy and personal safety concerns, we can not reveal our founder's personal information publically here, Judges please refer to "How do you plan to overcome these barriers?" part for our CEO's background. 

Qi, graduated from UConn with a dual degree in Computer Science and Cognitive Science.  She had four years of training in programming with specialties in software engineering and data analysis. She had 2-year experience in social media research which she gained proficient experience in using machine learning and artificial intelligence to solve social science questions. She also had proficient experience in software engineering. She had participated in several projects and had major contributions. The projects include machine learning personality classification tools for Twitter users, a data visualization website, and a Salesforce project managing tool.

Cecilia, graduated from UConn with a degree of Management Information System. She has experience as a project manager and frontline developer in both business and IT-focused projects during her study at school. She decided to defer her graduate school admission to John Hopkins University by taking the role of a team leader in the Azure Support Engineers team. She has past internship experiences in both the commercial banking and managerial consulting industry. As a former business analyst, she worked as an effective communicating bridge between developers and business owners. Cecilia is a great team organizer. Her effort as an empathetic intermediator and effective communicating person cannot be neglected in an ambitious tech-oriented team.

We believe and thrive in diversity. In an industry (AI/tech) that has long been dominated by white males, as a team that is made of women of color, our existence and story is empowering. 

The most important value of Alecto is compassion, we do not only integrate compassion into our product design, but also make sure our team adheres to that value. The core team is made of 3 women of color, but we are open for everyone to join, as long as they share our value of compassion. Outside of our core team, we also have contributors who came from the LGBTQ community.  

Alecto is beyond an AI product, it's a movement against digital abuse and algorithm bias. We aim to create positive impacts globally. Solve's network could help us scale and provide us with valuable advice. 

We hope to seek more capable and compassionate professionals with solid backgrounds in machine learning to join our team, and we hope MIT solve can help us make connections with them. 

In terms of partnerships, we hope to establish connection with organizations that shares our most important value (compassion) to help us grow and provide support for our users. 

We are looking for skilled professionals with a solid background in machine learning to help us scale the product. We also want to gather more technical specialties to help us with app development and UI/UX design. 

Since we aim at building our own online community, we are looking for people with marketing background and new media operation.  

Additionally, we wish to seek legal advice in terms of global regulation on facial recognition technology. 

We are open to partner with any organization within MIT solve's network 

Mental health is a very important part of health and well-being in general. Online image abuse victims oftentimes have to endure great mental distress, our founder is a revenge porn survivor, when she first found out about this brutal violation of her rights, she tried to commit suicide. By providing victims (or potential victims) with access to tools to defend their online identity and provide a community of support, we are addressing mental health issues for our users. We also want to provide several tools to our users for spiritual growth and development in order to help them better discover their inner selves. 

We believe that it is crucial to educate the public at a young age to help them build up awareness  of personal information security and privacy. 

We also want to plant the seed of compassion and goodness in young kids' hearts through supportive spiritual education. 

Our goal is to help as many digital abuse victims as we could, but due to operation cost concerns (human resources, server, tech, etc), we must be a for-profit company. Although our core business model is for-profit, we plan to include a non-profit aspect in our company. We plan to provide our service for free for victims who could not afford it. 

Alecto not only helps digital abuse victims, we are also working towards a future where algorithm bias based on race or gender doesn't exist. Our mission is to create an antiracist and equitable future for all (not only in the U.S., but also around the world). Advance technology is the tool to help us get there. 

Alecto not only provide AI graphic searching tool to the public, but also aim at building a compassionate and safe online community. Through this community, people are able to share their inner stories, provide mental supports to others, and gain knowledge on topics related to personal information security. 

Alecto is an unbiased facial recognition tool accessible for individuals to quickly search for infringing images online. Alecto will directly improve the lives of image abuse victims and people of color. By offering a compassionate, unbiased facial recognition tool that helps online image abuse victims to defend their privacy, we are making the online environment safer and more friendly, not only for communities of color, but also for everyone. By allowing individuals to safely train the machine learning model without jeopardizing their privacy, we are also making the facial recognition algorithm more human-centered, minimizing the risks of racial or gender bias.  

The team lead of Alecto is a revenge porn survivor, and a woman of color. Her personal experience helped her understand the pain points and needs of Alecto’s target population. Our core team is made of 3 women of color, in an industry that is dominated by white males (AI), our team offers a different approach and perspective. Alecto portraits an image of independent women of new era.  

Alecto is an unbiased facial recognition tool accessible for individuals to quickly search for infringing images online. Alecto will directly improve the lives of image abuse victims and people of color. By offering a compassionate, unbiased facial recognition tool that helps online image abuse victims to defend their privacy, we are making the online environment safer and more friendly, not only for communities of color, but also for everyone. By allowing individuals to safely train the machine learning model without jeopardizing their privacy, we are also making the facial recognition algorithm more human-centered, minimizing the risks of racial or gender bias.